Main

PKU due to phenylalanine hydroxylase deficiency is one of the most common inborn errors of metabolism. The phenylalanine hydroxylase gene could be mapped to chromosome 12q22-q24.1(1). Up to now more than 220 different mutant phenylalanine hydroxylese alleles have been described(2). Classification of hyperphenylalaninemia is done by serum phenylalanine levels under free diet: above 1200 μmol/L(“classical PKU”), between 600 and 1200 μmol/L (“mild PKU”), or below 600 μmol/L (“non-PKU hyperphenylalaninemia”),(3). Untreated PKU leads to mental and psychomotor retardation(1). Treatment consists of a protein/phenylalaninerestricted diet, supplemented with a phenylalanine-free amino acid mixture. Generally patients with non-PKU hyperphenylalaninemia were not treated, as normal outcome was suspected(4–6).

In spite of early and adequately dietary treatment, patients with mild or classical PKU show mild reduction of IQ(7–9), abnormalities of cerebral white matter(10, 11), and deficits of fine motor performances(12), sustained and selective attention(13, 14), as well as impaired “executive functions” depending on (pre-) frontal lobe functions. The last are discussed to be consequence of a central dopamine depletion(14–16). In addition neurologic deterioration was described in some adult patients, who stopped their diet for more than 10 y(17).

Most recently Costello et al.(18) investigated a group of untreated patients with mild PKU and non-PKU hyperphenylalaninemia (serum phenylalanine levels after the first test consistently below 900 μmol/L). At the age of 4 y these patients presented a lower IQ (mean = 103.6, SD = 18.0) compared with treated patients (mean = 110.0, SD = 18.0)(18). As a consequence of these results, diet policy for patients with hyperphenylalaninemia was changed, and a lifelong diet with serum phenylalanine levels below 400 μmol/L was recommended even for those patients with serum phenylalanine levels consistently below 600 μmol/L under free diet(18, 19).

Up to now in most centers patients with non-PKU hyperphenylalaninemia are not treated. The intellectual, neurologic, and neuropsychologic outcome of untreated adolescent and adult patients with non-PKU hyperphenylalaninemia has not so far been studied systematically. This study was initiated to answer the question whether patients with non-PKU hyperphenylalaninemia are at risk for developing intellectual, neurologic, and neuropsychologic impairment without dietary treatment.

METHODS

Subjects. Twenty-eight of a total of 40 untreated adolescents and adults with non-PKU hyperphenylalaninemia of four German treatment centers(Düsseldorf, Hannover, Heidelberg, and Münster) agreed to take part in our study (Table 1). Twelve subjects were not willing to participate, because they expected no benefit for themselves. These subjects did not significantly differ from the participating subjects in their socioeconomic status and their school career (information by telephone). For four patients we were not able to find a comparable healthy control subject. Therefore these four patients were partly excluded from evaluation (see subsamples in Table 1). Healthy control subjects were employees of the clinic and their relatives who were selected according to their age, sex, and socioeconomic status. Subjects with untreated non-PKU hyperphenylalaninemia were defined by blood pheylalanine levels consistently below 600 μmol/L in the screening test, sporadic tests throughout life, as well as the mean serum phenylalanine level of four determinations at 2-wk intervals during the study (= mean study level). According to the classification of hyperphenylalaninemia, mentioned above(3), patients with serum phenylalanine levels above 600μmol/L in any of these variables (screening test, sporadic tests throughout life, four determination at 2-wk intervals during the study period) were excluded from the study. Serum phenylalanine levels of the screening test and mean study levels were highly correlated (see “Results”). Both were used to differentiate between patients with serum phenylanine levels below versus above 360 μmol/L. This limit of 360 μmol/L was taken because the German Pediatric Society recommended serum phenylalanine levels below 360 μmol/L for patients with mild or classical PKU up to the age of 10 y(11). For 25 patients blood phenylalanine levels of the newborn screening test according to Guthrie (Guthrie bacterial inhibition assay) were available.

Table 1 Description of study groups
Full size table

Three elder subjects included in the sample were born before the general screening program was introduced. Blood phenylalanine levels of the newborn screening are strongly correlated with the predicted level of phenylalanine hydroxylase activity, based on expression analysis(2, 20). In addition seven patients were characterized by a Blascovics load at the age of 4-6 mo(21). A mean lifetime serum phenylalanine level was not computed because serum phenylalanine checks were done only sporadically(Table 2). Socioeconomic status was defined with the help of the variables family income, degree of father, and job of father as described earlier(22).

Table 2 Serum phenylalanine levels
Full size table

Mutation analysis. All 13 exons and flanking intronic sequences of the phenylalanine hydroxylase gene from 28 non-PKU hyperphenylketonuric subjects were amplified by polymerase chain reaction and scanned for the presence of mutations by denaturant grading gel electrophoresis according to previously described procedures(23, 24). All samples displaying an altered electrophoretic band pattern were subjected to direct sequencing.

Instruments. The assessment program consisted of tests that had revealed significant deficits in early treated patients with mild or classical PKU(8, 11–16). All 28 patients were examined for clinical-neurologic outcome, IQ, school performance, and job career. A subsample of 24 patients and 24 healthy control subjects, matched for age, sex, IQ, and socioeconomic status were tested for selective attention (CWIT = Stroop task), sustained attention (DPE from the Sonneville Visual Attention Task), and for their “executive functions” (Tower of Hanoi Planning Task), as well as for fine motor performances (MPT). All patients and controls were right-handed.

Cranial MRI was obtained in 10 of the patients and 10 of the healthy control subjects, matched for age and sex. Clinicalneurologic examination was done by the same neurologist for patients and control subjects.

IQ was tested by the German version of the WAIS-R(25). Two 14 y-old-boys (one subject with non-PKU hyperphenylalaninemia and one healthy control subject) were tested with the WISC-R. Internal consistency of the WAIS-R was r = 0.96. The test was standardized in 1991 with a norm population (n = 1000) homogeneous for age and sex (age range: 16-34 y).

Details concerning school performance and job career were provided by the mothers' ratings of the 28 patients and a total of 28 healthy siblings. Age at start of schooling, type of school, and repetition of classes were taken as criteria evaluating school career. For job career, the type of job was regarded. Healthy siblings themselves were not examined in our study, because most of them were not willing to participate.

Fine motor performances were assessed by the MPT(26). The computer-based test by Schoppe consists of six subtests, measuring steadiness, precision, speed, and visual-motor coordination of finger-hand-arm movements. The test results of only the dominant right hand were described because left-hand results showed no significant differences between the two groups. Weglage et al.(12) reported significant deficits on the MPT in early treated adolescents with PKU.

Sustained attention was measured by the DPE of the Sonneville Visual Attention Task(27). This test is a self-paced continuous performance task. The total number of 600 reactions is divided into 50 series, each with 12 signals. For each series, reactions times of the 12 reactions are cumulated. The mean reaction time of these 50 series (mean series) indicates the overall level of performance, whereas the SD of this variable mean series is the measure of the stability of performance. Schmidt et al.(13) reported significant deficits on sustained attention in early treated PKU patients using the DPE. Selective attention was assessed by the CWIT according to Stroop(28). The CWIT consists of three parts: reading of color words, color naming, and the interference task. The interference is quantified in terms of the increase in reaction time compared with reading of color words and color naming. Positron emission tomography revealed a frontal lobe activation in healthy adults during the Stroop task(29). In analyzing Stroop test results of patients with frontal lesions, the CWIT was demonstrated to be dependent on frontal lobe functions(30). Early treated patients with PKU showed significant deficits on the CWIT(14, 15).

“Executive functions” were examined by the Tower of Hanoi Planning Task(16). This task evaluates the ability to plan and execute a sequence of moves that transform an initial configuration of disks into a goal state that duplicates the experimenter's disk configuration. Five problems varying in initial disk configuration and difficulty were administered. Deficits in this type of planning behavior have been demonstrated in frontal damaged adults(30) and early treated children with PKU(16). The evaluation was done as described by Welsh et al.(16).

Ten patients and 10 healthy control subjects, matched for age and sex, were examined with a superconducting 1.5 T magnet (Magneton, Siemens) using spin-echo pulse sequences. T1-weighted and proton density-weighted images [transmission time (TR) 2500, echo time (TE) 15/90] were acquired in the axial plane. In addition coronal T2-weighted images (TR 600, TE 1.5) were obtained. Five subjects with non-PKU hyperphenylalaninemia had mean study levels below 300 μmol/L (252± 60 μmol/L), and in five subjects the level was above 300 μmol/L(528 ± 126 μmol/L).

Data analysis. Parametric tests (one-sided t test, correlation according to Pearson) and nonparametric tests (χ2 test) were used for statistical evaluation of the data. The α error was corrected according to Bonferroni.

RESULTS

Genotypes and corresponding biochemical phenotypes. The genotypes and corresponding serum phenylalanine levels for the 28 patients with non-PKU hyperphenylalaninemia are given in Table 3. The present investigation supports previous suggestions that non-PKU hyperphenylalaninemia is caused by compound heterozygosity for a PKU mutation and a non-PKU hyperphenylalaninemia mutation(24). The PKU mutations are indicated as allele I and the non-PKU hyperphenylalaninemia mutations as allele II (Table 3). The mutation detection rate was 80%, which is the lowest mutation detection rate obtained by the present method. Sequence analysis of the uncharacterized chromosomes did not reveal any DNA sequence alterations in the 13 exon regions, and the gene defect reponsible for phenylalanine hydroxylase deficiency on these chromosomes is thus probably situated outside the region of investigation,i.e. at the polyadenylation sites or in the large intronic sequences. The A403V mutation happens to be the most prevalent hyperphenylalaninemia mutation in this sample, representing 24% of all mutations discovered. One novel mutation was discovered, i.e. Q20X(CAG-TAG). One patient (no. 19) is homozygous for the A403V mutation, indicating that this mutation codes non-PKU hyperphenylalaninemia. The mutations A403V, V245A, E390G, and S87R have previously been designated associated with the non-PKU hyperphenylalaninemia phenotype as they have been found in non-PKU hyperphenylalaninemic individuals harboring a mutation on the other chromosome known to completely abolish phenylalanine hydroxylase activity(23, 24). Two of the subjects (nos. 11 and 25) have inherited two mutations (L48S and Y414C), each of which results in mild PKU when expressed in combination with a mutation that completely abolishes phenylalanine hydroxylase activity.

Table 3 Genotypes and serum phenylalanine levels of 28 patients
Full size table

Clinical-neurologic outcome. Clinical-neurologic examination revealed a mild rest and intention tremor in one subject with non-PKU hyperphenylalaninemia and a clumsy diadochokinesis in another. One control subject presented a mild rest and intention tremor. Physical development was normal in all subjects with non-PKU hyperphenylalaninemia(31).

Intelligence. The 28 subjects with non-PKU hyperphenylalaninemia attained a normal IQ. Verbal IQ and performance IQ did not significantly differ (Table 4).

Table 4 Intelligence scores
Full size table

School performance. The 28 subjects with non-PKU hyperphenylalaninemia and their 28 siblings did not significantly differ with regard to age at start of schooling, types of schools, and repetition of classes (Table 5).

Table 5 School performance
Full size table

Job career. Subjects with non-PKU hyperphenylalaninemia and siblings did not significantly differ with respect to their jobs(Table 6).

Table 6 Job career
Full size table

Sustained attention. Levels of performance as well as the stability of performance on the DPEs of the Sonneville Visual Attention Task were not significantly different for subjects with non-PKU hyperphenylalaninemia and control subjects (Table 7).

Table 7 Sustained attention DPE
Full size table

Selective attention. The CWIT, which depends on frontal lobe functions, revealed no significant difference between subjects with non-PKU hyperphenylalaninemia and control subjects (Table 8).

Table 8 Selective attention (CWIT, Stroop task)
Full size table

Executive functions.Using the Tower of Hanoi Planning Task patients showed no deficits of executive functions, which again are discussed to depend on prefrontal lobe functions (Table 9).

Table 9 Executive functions (Tower of Hanoi planning task)
Full size table

Fine motor performance. The MPT revealed no significant fine motor deficits in subjects with non-PKU hyperphenylalaninemia(Table 10).

Table 10 MPT-dominant right hand
Full size table

MRI. The blind evaluation of subject and control MRI scans revealed no abnormalities in any case, especially no alterations of cerebral white matter.

Correlational findings. The phenylalanine levels of the screening test and the mean study levels were significantly correlated(r = 0.64, p < 0.001). Subjects with blood phenylalanine levels (screening test and mean study levels) below 360μmol/L reached a significantly lower IQ compared with those with blood phenylalanine levels between 360 and 600 μmol/L (mean = 100.1, SD = 16.2,n = 16 versus mean = 105.2, SD = 15.2, n = 12)(t test: p < 0.05). With this exception subjects from both subgroups did not show any significant difference in their results(t tests: NS). The 10 patients with serum phenylalanine levels(screening and mean study levels) above 360 μmol/L and the 10 healthy control subjects, matched for age, sex, IQ, and socioeconomic status, did not significantly differ in any of the test results (DPE, CWIT, Tower of Hanoi, MPT; t tests: NS). Correlational analysis according to Pearson revealed no significant relation between serum phenylalanine levels (screening test and mean study level) and test results.

DISCUSSION

Results presented indicate that adolescent and adult subjects with untreated non-PKU hyperphenylalaninemia seem to be not at risk for developing remarkable intellectual, neurologic, and neuropsychologic deficits, as well as abnormalities of cerebral white matter, as described for patients with mild or classical PKU(3, 11–16). Further studies with larger samples of untreated subjects and phenylalanine levels especially in the controversial discussed range between 360 and 600 μmol/L will have to confirm our preliminary results. But our results were supported by findings of Levy et al.(32) who reported normal IQs in 13 4-y-old patients with non-PKU hyperphenylalaninemia that were comparable to the IQs of their unaffected siblings(4). In another study of Levy and Waisbren(5) eight women with non-PKU hyperphenylalaninemia reached a normal mean IQ. In a study on maternal mild hyperphenylalaninemia Levy et al.(32) found a normal median IQ (105) in 37 of 86 women for whom IQ test results were available. Patients with phenylalanine levels below 400 μmol/L had a higher median IQ (108) compared with patients with phenylalanine levels below this level (100), but this difference was not significant(32). Waisbren et al.(6) found no demonstrative intellectual deficits or abnormalities in personality in a group of 12 patients with untreated mild PKU (n = 6) and non-PKU hyperphenylalaninemia (n = 6). Unfortunately nothing is known about the socioeconomic status of the patients in these studies.

In our study, results of different tests were not significantly influenced by blood phenylalanine levels. In comparison with subjects with blood phenylalanine levels below 360 μmol/L as well as in comparison with healthy control subjects, subjects with blood phenylalanine levels between 360 and 600μmol/L (screening test and mean study level) did not reach significantly poorer results. (The higher mean IQ of the patient group with serum phenylalanine levels above 360 μmol/L might be interpreted as a result of pure chance.) Surprisingly this also holds true for the CWIT and the Tower of Hanoi, both depending on frontal lobe functions. Diamond(15) and Welsh et al.(16) reported reversible deficits using these tests in preschool subjects with serum phenylalanine levels between 360 and 600μmol/L. Using the CWIT in 10-y-old patients with early treated PKU we were able to confirm these results(14). Deficits were mild and school performance as well as the IQ of the patients were not impaired(14). In contrast, preliminary results of seven adult patients with classical PKU on the CWIT were found to be normal(33). As the maturation of the frontal lobes seems to be finished during early adolescence these deficits might be an age-related temporary phenomenon(34).

Our results contrast to the “British Medical Research Council Working Party on Phenylketonuria” which recommended a lifelong dietary treatment for all patients with serum phenylalanine levels above 400 μmol/L(19). This recommendation is based on a study of Costello et al. (British Register on Phenylketonuria)(18). The limit of 400 μmol/L was justified by the fact that patients with mild PKU and non-PKU hyperphenylalaninemia (serum phenylalanine levels consistently below 400 μmol/L) reched an above normal IQ(18).

In contrast, deficits in IQ were reported for a group of 4-y-old patients with untreated mild PKU and non-PKU hyperphenylalaninemia(18). Unfortunately subjects with serum phenylalanine levels above 600 μmol/L in the screening test were included in the study, if subsequent values were consistently below 600 μmol/L. In these subjects protein restriction alone was not counted as treatment. The study does not differentiate between subjects with non-PKU hyperphenylalaninemia and mild PKU, but recommends dietary treatment for all patients with serum phenylalanine levels above 400 μmol/L. In the study of Costello et al.(18) the IQ was measured at the age of 4 y. IQ test results of 4-y-old subjects are not as reliable as they are in adolescence and adulthood(7). Furthermore, Costello et al.(18) used an antiquated IQ test(Stanford-Binet test, short form L-M, standardized 1932, 1960 revision). In addition, Costello et al.(18) concluded that every 100 μmol/L increase in mean phenylalanine concentration (throughout life) leads to a decrease of approximately 7 IQ points. This conclusion is based on a linear regression. The simplified model does not take into account that elevated serum phenylalanine levels early in life seem to be more critical for intellectual development than such levels later in life(7). Furthermore, it is not clear whether such a simple linear relation exists for all patients with classical, mild, or non-PKU hyperphenylalaninemia. We believe that the results of our study are of great social import. About 1 of 10 000 newbrons in the United States and Europe suffers from hyperphenylalaninemia. Approximately 30% of these patients have non-PKU hyperphenylalaninemia(8). Compared with a normal nutrition, a phenylalanine-restricted diet in Europe (amino acid supplementation and special food) would additionally cost approximately $35 000 for one subject with non-PKU hyperphenylalaninemia up to the age of 15 y(35). A lifelong diet, as recommended by the British Medical Research Council Working Party on Phenylketonuria, would be even more expensive. In this calculation, costs of other medical care were not taken into account(35).

In addition, the treatment of the chronic disease hyperphenylalaninemia is burdensome, and the diet may lead to severe psychologic and social problems, including depressive mood, negative self-image, health worries, and social stigmatization as we described recently(22).

In conclusion, in the absence of any demonstrative effect, treatment is unlikely to be of significant benefit to subjects with non-PKU hyperphenylalaninemia. The subjects participating in this study were defined by phenylalanine concentrations below 600 μmol/L in the newborn screening, sporadical serum phenylalanine determinations throughout life, and in the four serum phenylalanine determinations in 2-wk intervals during the study. Therefore we are quite sure that these subjects really have non-PKU hyperphenylalaninemia and are not PKU patients with a delayed increase in blood phenylalanine concentration initially classified as non-PKU hyperphenylalaninemias and described most recently by Berlin et al.(36).

Nevertheless, problems of maternal PKU should be intensively discussed with non-PKU hyperphenylalaninemic female patients and their relatives, as their elevated serum phenylalanine levels during pregnancy may be associated with damage to the fetus(5, 32, 37).