By Keith F. Widaman, Ph.D.
The International Collaborative Study of Maternal Phenylketonuria (ICSMPKU) was designed to monitor and treat the pregnancies of women with phenylketonuria (PKU). Early research on untreated pregnancies of women with PKU documented an array of negative fetal outcomes, including mental retardation, microcephaly, congenital heart disease, low birthweight, and spontaneous abortion (Lenke, Levy, 1980). Each of these outcomes tended to show a dose-response curve, with higher levels of maternal phenylalanine (Phe) when on a regular diet associated with higher levels of negative outcomes for offspring. For example, 73 percent of the offspring of mothers with classical PKU (Phe
<=1200 µmol/L) had microcephaly, whereas 68 percent of offspring of mothers with moderate PKU (1200 µmol/L > Phe
<=900 µmol/L), 35 percent of offspring of mothers with mild PKU (900 µmol/L > Phe
<=600 µmol/L), and 24 percent of offspring of mothers with mild hyperphenylalaninemia (600 µmol/L > Phe
<=180 µmol/L) had microcephaly.
A primary goal of the ICSMPKU was to encourage pregnant women with PKU to remain on a low Phe diet and thereby to maintain low Phe levels during pregnancy. This effort was, as an overall intervention, highly successful. In these treated pregnancies, only 23 percent of offspring of mothers with classical PKU, 8 percent of offspring of mothers with moderate PKU, 10 percent of offspring of mothers with mild PKU, and 5 percent of offspring of mothers with mild hyperphenylalaninemia had microcephaly (Koch, Friedman, Azen, et al., 1999).
However, large differences in average Phe levels during pregnancy were still apparent in the ICSMPKU study sample. Some mothers were able to maintain average Phe levels of less than 180 µmol/L throughout pregnancy, whereas other mothers had average Phe levels of more than 1200 µmol/L. The potential reasons for this large range in average Phe levels during pregnancy are numerous, including the poor taste of low-Phe diet food and the failure of mothers to comprehend the potential danger to the fetus of high maternal Phe levels. Another major factor was the mother’s Phe level when on a regular diet. Mothers with classical PKU found it much more difficult to maintain low Phe levels than did mothers with less severe forms of PKU.
The large range of individual differences in average Phe levels during pregnancy presented us with an opportunity to explore the causes and consequences of average Phe levels. The results presented here have not yet appeared in a peer-reviewed journal, but they are key for describing (1) the amount and extent of Phe exposure necessary to produce negative effects on offspring; (2) the causes of, or the variables that influence, maternal Phe levels during pregnancy; (3) the consequences of Phe level during pregnancy; and (4) what remains unknown about the effects of prenatal exposure to high levels of Phe on developmental outcomes in offspring.
Prior research was equivocal about the level of Phe that could be tolerated by mothers with no ill effects on the fetus, so several important outcomes in offspring were measured in the ICSMPKU, including congenital heart disease and intelligence quotient (IQ). A consistent finding of current analyses is that average Phe exposure of up to approximately 360 µmol/L has no effect on offspring IQ. But for exposure above that level, infant IQ showed a 3- to 3.5-point drop for every additional 60 µmol/L. Thus, infants who were exposed during pregnancy to an average Phe of 600 µmol/L (the upper limit defining hyperphenylalaninemia) had a mean infant IQ between 12 and 14 points lower than did infants exposed to average Phe levels of 360 µmol/L or less.
The timing of Phe exposure during pregnancy is also a crucial variable. Average Phe levels during the first 2 months of gestation had rather low levels of correlation with infant and child IQ (below .20). But from the third month of gestation on, the correlations between average Phe levels and infant and child IQ rose to the .30 to .40 range. These results are consistent with the hypothesis that Phe exposure has damaging effects on the fetal systems undergoing greatest growth during that period. Because the greatest amount of cortical development occurs during the latter two trimesters of pregnancy, average Phe levels during those periods correlate most strongly with intelligence, which depends on adequate development of cortical areas.
However, for other outcomes, such as congenital heart problems, average Phe levels during the first 2 months of gestation appear to be crucial, since the heart is undergoing rapid development during this period. These findings lead to a general conclusion: The amount and timing of Phe exposure during gestation that lead to negative effects are likely to differ across type of outcome. That is, certain outcomes may be affected by high Phe exposure during early pregnancy, whereas other outcomes are affected only by high Phe exposure during later pregnancy. In addition, the amount of Phe exposure necessary for negative outcomes may vary; therefore, determining the levels of Phe exposure that lead to teratogenic effects is an important research endeavor.
Given our results showing that average Phe levels during pregnancy were related in a nonlinear fashion to infant and child intelligence, we turned our attention to variables that influence Phe levels during pregnancy. In work to date, we have documented the influences of maternal education, intelligence, and socioeconomic status on average Phe levels during pregnancy. Of these three variables, maternal intelligence is the strongest (
<= = -.47), whereas maternal education and socioeconomic status are moderate in their effects (
<=s around -.30). All of these effects are in the predicted direction: Higher levels of maternal intelligence, education, and socioeconomic status lead to lower levels of average Phe during pregnancy. These results suggest that practitioners should concentrate their efforts to ensure compliance with the low-Phe diet regimen in pregnant women with lower levels of IQ and education.
The consequences of high Phe levels during pregnancy for infant and child intelligence are quite strong. To assess offspring outcomes in the domain of mental abilities, the ICSMPKU used the Bayley Scales of Infant Development at 1 and 2 years of age, the McCarthy Scales of Children’s Abilities at 4 years of age, and the Wechsler Intelligence Scale for Children–III (WISC-III) at 7 years of age.In predicting the Bayley Mental Development Index (MDI) at 1 year of age, average Phe exposure was the strongest predictor (
<= = -.39); other predictors had smaller regression weights (
<=s < |.20|).Using the Bayley MDI at 2 years as the criterion, average Phe exposure was again the strongest predictor (
<== -.46); other predictors had smaller regression weights (
<=s < |.22|).
A similar trend held for the McCarthy Scale General Cognitive Index (GCI) and the WISC-III Full Scale IQ (FIQ).Average Phe exposure during gestation was the strongest predictor, whereas other predictors had rather small regression weights.The relationship between average Phe exposure during gestation and intelligence got stronger as age increased.For example, the effect of average Phe exposure on the GCI at 4 years was
<= = -.53, and on the FIQ at 7 years was
<= = -.59.These results run counter to those found in most developmental studies, where the strength of effect of a predictor decreases as the length of time between measurements increases.
Much remains unknown about the effects of prenatal exposure to high levels of Phe on later developmental outcomes.Analyses of the outcomes at 1, 2, 4, and 7 years converged on the estimate of average Phe exposure of 360 µmol/L as the point at which teratogenic effects on offspring begin to appear.We currently have no firm estimates of the amount of Phe exposure that can be tolerated before teratogenic effects appear in other key outcomes.It appears that exposure to high levels of Phe in the first 2 months of gestation may contribute to congenital heart problems, and that exposure to high levels of Phe in the latter two trimesters of gestation affects intelligence, but the timing of Phe exposure for other developmental outcomes is little understood.Furthermore, we can only speculate on the effects of prenatal exposure to high levels of Phe for intellectual development during adolescence and adulthood.The most prudent guess would be that children exposed to high Phe levels during gestation will continue to show pronounced deficits during adolescence and adulthood, especially given the trend of stronger effects of prenatal Phe exposure on intellectual measures at later ages.Whether behavioral or biomedical treatments can alleviate the negative effects of exposure to high Phe levels during gestation is a topic to be pursued in future research.We can only supply a motivation for such research by documenting the pervasive negative effects of prenatal exposure to high levels of Phe.
- Koch R, Friedman E, Azen C, Hanley W, Levy H, Matalon R, et al. The International Collaborative Study of Maternal Phenylketonuria status report 1998. MRDD Res Rev1999;5:117-21.
- Lenke RR, Levy HL. Maternal phenylketonuria and hyperphenylalaninemia. An international survey of the outcome of untreated and treated pregnancies. N Eng J Med 1980;303:1202-8.
- Widaman KF. Process of analyses of data: benefits and costs associated with collaborative studies. MRDD Res Rev1999;5:155-61.