Cognition and Executive Functioning

Children exposed prenatally to alcohol typically have intelligence quotient (IQ) scores in the low average to borderline range. Children with more dysmorphic features tend to have lower IQ scores, but cognitive problems are not limited to this group. Of those with FAS, around 25% to 50% will experience severe cognitive delay (IQ score <70), with the pooled prevalence of cognitive disability being 97 times higher in infants with FAS than the general population. In addition to global cognitive impairment, deficits in executive function and memory are also evident on neuropsychological testing. Executive deficits include problems with planning and organization, cognitive flexibility/set shifting, working memory, and behavioral inhibition, with parents reporting the greatest difficulty with inhibitory control and problem solving. In terms of memory deficits, problems encoding or memorizing information appear more prominent than problems with recall. Not surprisingly, given this constellation of cognitive impairments, learning problems are very common at school, even after controlling for IQ.

Language Development

Children with FASD are characterized by delays in the acquisition of language and understanding of spoken language. Common difficulties include word comprehension, naming ability, phonological processing, speech production, and articulation errors, resulting in poorer performance on tests of both receptive and expressive language development.

Behavior and Regulatory Problems

Virtually all infants with FASD exhibit serious attentional and behavioral problems, with 70% subsequently meeting clinical criteria for a diagnosis of ADHD. The next most common condition is oppositional defiant/conduct disorder, with children with FASD showing less guilt after misbehaving, less behavioral maturity for their age, and higher levels of antisocial behavior, including cruelty and stealing. Longer-term substance abuse and mood disorders, such as anxiety and depression, are also common. For example, relative to the general population, adults with FASD have more hospital admissions for alcohol abuse (9% vs. 2%) and psychiatric disorders (33% vs. 5%), and are also more likely to be prescribed psychotropic medications (57% vs. 27%).

Motor and Visual Function

Finally, deficits in both fine and gross motor development have been reported in children with FASD and include tremors, weak grasp, poor hand-eye coordination, and impaired postural balance. Visuospatial performance indicates poorer saccadic control, which results in the processing of visual stimuli in a disorganized way.

Fetal Blood Flow

Maternal alcohol exposure has a variable impact on uterine blood flow, depending on the gestation of the fetus as well as the pattern of exposure. However, there is converging evidence to suggest that alcohol alters the development of new blood vessels and vascular remodeling, which are both essential to normal uteroplacental circulation during gestation. There is also evidence that alcohol may alter cerebral oxygen and glucose consumption in the fetus near term gestation, but these effects do not appear earlier in gestation. Further studies are needed to better understand how alcohol consumption affects uteroplacental hemodynamics during different maturational periods of pregnancy.

Fetal Malnutrition

The quality of maternal nutrition as well as the direct physiologic effects of alcohol on the fetus are also likely involved in the pathogenesis of FASD. Specifically, poor maternal nutritional status and vitamin deficiencies are common comorbidities of chronic alcoholism. Alcohol also interferes directly with the absorption, digestion, and utilization of nutrients. Retinol (vitamin A), folate, and zinc, three nutrients that are important for fetal brain development, have been shown to be directly affected by maternal alcohol use. First, retinoic acid, the oxidized form of retinol, plays a pivotal role in the development of the nervous system as well as limb morphogenesis. Alcohol competitively inhibits the oxidation of retinol in the liver. In mouse models, deficiencies in retinoic acid early in gestation alter the expression of the sonic hedgehog gene, resulting in the classic craniofacial and corpus callosum abnormalities characterizing severe FAS. Second, folic acid has numerous roles in the developing nervous system, especially neural tube closure (see Chapter 1 ). Alcohol interferes with folic acid absorption, inhibits its metabolism, and increases excretion. Clinical research has shown a significant reduction in fetal-to-maternal folate ratios associated with chronic maternal alcohol use. Third, zinc is necessary for neurogenesis, neuronal migration, and synaptogenesis. Alcohol induces the zinc binding protein metallothionein in the maternal liver, which sequesters zinc and results in fetal deficiency. Although other primary and secondary nutritional deficiencies may exist in the fetuses of mothers who consume alcohol during pregnancy, the relative importance of these perturbations in the overall pathogenesis of FASD requires additional investigation.

Molecular Effects

Major contributors to the cascade of developmental damage associated with fetal alcohol exposure include (1) excessive cell death (apoptosis), (2) deficient cell proliferation, (3) impaired cell migration, and (4) altered differentiation. In experimental models, neuronal and oligodendrocyte degeneration is a prominent feature of prenatal alcohol toxicity. In particular, alcohol induces two types of cell death in the developing fetal brain. These include apoptosis and necrosis. Necrosis appears to represent a minority of cell death associated with alcohol exposure, typically occurring following binge drinking or alcohol withdrawal. Even low levels of alcohol exposure early in gestation appear to promote apoptosis, via elevation of phospholipase C activity, promotion of intracellular calcium transit, and repression of the transcriptional effector, β-catenin. This is an abnormal process during this stage of development. Later in gestation, apoptosis naturally occurs in approximately one-third of postmitotic neurons. However, in experimental models, alcohol exposure may increase the extent of apoptosis, most likely through N -methyl- d -aspartate (NMDA) receptor blockade and hyperactivation of γ-aminobutyric acid type A (GABA _A ) receptors.

Experimental models of fetal alcohol exposure during the equivalent of the human second trimester demonstrate a profound impact on neuronal proliferation, migration, and differentiation. Organization of neural circuitry relies on ongoing electrochemical activity, allowing firing neurons to locate and synapse with other cells as part of activity-dependent network formation. Neurosuppression via NMDA antagonism/GABA _A agonism depresses electrochemical activity, suppressing neurogenesis and triggering apoptotic neuronal death in the developing brain. In addition, alcohol impairs the function of insulin-like growth factor receptors, potentially via inhibition of cyclins and cyclin-dependent kinases. This inhibition may help explain delays in progression through the neuronal cell cycle that have been observed in experimental models. Alcohol induces errors in neuronal migration, perhaps through disruption of L1 cell adhesion molecule and interference with glial fibers. Alcohol phosphorylates the cytoplasmic domain of L1 cell adhesion molecule, thus altering the conformation and function of the extracellular domain. These changes not only affect neuronal migration, but also promote aberrant dendritic morphology.

Genetic and Epigenetic Alterations

Maternal and fetal genetics play a prominent role in susceptibility to FASD. Alcohol is metabolized in the liver to acetaldehyde by alcohol dehydrogenase (ADH). Functional polymorphisms in the locus encoding the beta subunit of the Class I ADH (ADH1B) alter the rate of alcohol metabolism. Alcohol is more rapidly metabolized in individuals with the ADH1B*3 allele (15% to 20% of African Americans) compared with individuals with the more common ADH1B*1 allele. This finding appears to have functional significance because prenatal alcohol exposure has been associated with increased attention problems and externalizing behavior in adolescents born to mothers with two ADH1B*1 alleles. In contrast, these differences were not observed in adolescents whose mothers had at least one ADH1B*3 allele.

Fetal genetic differences may also play an important role in outcome . After prenatal alcohol exposure, monozygotic twins are more concordant in outcome than dizygotic twins. Neuronal nitric oxide synthase (nNOS) and oxyguanine glycosylase 1 (OGG1) protect the developing brain from injury induced by alcohol. Mice homozygous for null nNOS or OGG1 have more severe neuronal damage and functional deficiencies following alcohol exposure compared with wild-type mice. In contrast, homozygous mutation of the tissue plasminogen activator and Bax genes are protective against the effects of alcohol on fetal brain development. Further research is required to determine the array of genes that influence the spectrum of fetal alcohol disorders.

Changes in gene expression after prenatal alcohol exposure represent a relatively new field of study. However, there is evidence to suggest that alcohol alters the sequence of genes involved in methylation, chromatin remodeling, protein synthesis, and mRNA splicing. Alcohol also promotes epigenetic changes, including alterations in methylation and acetylation, that give rise to phenotypic changes in animal models. As noted earlier, there is also evidence that epigenetic modifications associated with prenatal alcohol exposure may be transgenerational, placing subsequent generations at increased risk of alcohol dependence. Further study of the mechanisms underlying these important transgenerational effects is needed.

Maternal-Fetal Effects

In addition to the risks of major congenital malformations (see later), a recent meta-analysis of 38 studies across both low and high income countries found that infants born to pregnant women who were treated with AEDs for epilepsy had increased odds of preterm (<37 weeks gestation) birth (OR = 1.16, 95% CI 1.01 to 1.34) and fetal growth restriction (OR = 1.26, 95% CI 1.20 to 1.33). Fetal and neonatal mortality were not increased. However, this finding is contrasted by recent observational data suggesting a higher incidence of intrauterine death associated with maternal AED polytherapy. Microcephaly has also been noted in about 12% of cases. For example, a prospective longitudinal study of 329 pregnant women with epilepsy reported an increased risk of microcephaly at birth and at 12 months of age (12%), but which normalized by 24 months. More research is needed to examine other potential neonatal neurobehavioral outcomes, as well as the extent to which these risks vary by AED, dose, epilepsy type, and therapy regimen (monodrug vs. polydrug exposure).

Major Congenital Malformations

AED exposure in pregnancy results in a constellation of fetal and developmental effects that range in severity from major congenital malformations to subtle variations in normal development. Early, large-scale retrospective studies in the 1970s and 1980s were the first to document the teratogenic effects of AEDs on the developing fetus. Findings suggested a more than twofold increased risk of major malformations in offspring of epileptic women on AEDs versus those not on medication or nonepileptic women. Subsequent studies in the 1990s supported this initial work. These studies focused largely on fetal valproate syndrome and fetal hydantoin syndrome ; the latter is a syndrome associated with phenytoin exposure or structurally similar agents, such as phenobarbital, carbamazepine, and oxcarbazepine. The clinical features of these syndromes are reviewed in Table 38.6 and Figs. 38.6 and 38.7 .

Fetal hydantoin syndrome: clinical appearance.

(A and B) Two different infants with the syndrome; note the similar facial appearance, especially the broad, depressed nasal bridge, the widely spaced eyes, the epicanthal folds, the low hairline at the forehead, and the short neck. (C) Hand of an affected newborn. Note hypoplasia of nails and distal phalanges. (D) Hand of an affected child at 1 year of age; some nail growth is evident.

(From Hill RM, Verniaud WM, Horning MG, McCulley LB, Morgan NF. Infants exposed in utero to antiepileptic drugs. Am J Dis Child . 1974;127:645–653.)

Fetal valproate syndrome, clinical appearance.

Infant (10 months old) whose gestation was complicated by administration of 2000 mg of valproate daily to the mother: note the microbrachycephaly, apparent hypertelorism, broad flat nasal bridge, small mouth with thin lips (A), and thin, tapering fingers (B).

(From Schorry EK, Oppenheimer SG, Saal HM. Valproate embryopathy: clinical and cognitive profile in 5 siblings. Am J Med Genet A . 2005;133:202–206.)

More recently, given the increased use of AEDs and a growing awareness of adverse effects on the cognition and behavior of children born to treated women, a number of new important data sources have been established. These include (1) large registries of pregnant epileptic women in North America, Scandinavia, the United Kingdom, India, and Australia ; (2) pharmaceutical company registries for specific drugs; (3) the European Surveillance of Congenital Anomalies (EUROCAT) database across 14 countries ; (4) the US National Birth Defects Prevention Study (NBDPS) ; and (5) prospective cohort studies of pregnant women treated with AEDs and their infants. Although not without methodological challenges, including limited information on newer AEDs, insufficient follow-up data, and variable methods used, these efforts have helped advance understanding of the neonatal and longer-term effects of AED exposure during pregnancy.

It is clear from this work that several factors have an influence on the teratogenic effects of AEDs ( Table 38.7 ). First, fetal effects differ for each AED . Valproate appears to be the most teratogenic, followed by phenobarbital and topiramate that carry moderate risks. Newer drugs, at least based on current evidence, appear to be generally less harmful, although not all are completely risk-free ( Table 38.8 ). Second, the dose of drug has been found to influence outcome. Higher maternal doses tend to carry more risk. Third, the timing of AED exposure during pregnancy is also important. This fact poses challenges, given that neural tube closure occurs between weeks 3 and 4 of gestation and systemic organogenesis between weeks 4 and 10. This early period of vulnerability is a time when women are often unaware of their pregnancy. Fourth, the type of epilepsy can further complicate outcomes for the fetus. Specifically, breakthrough seizures during pregnancy may result directly in fetal hypoxia, and consequent falls may increase the risk of premature labor and fetal death. Fifth, the presence of other comorbid conditions , such as psychiatric and endocrine disorders, and the potential for polypharmacy (e.g., antidepressants) may further increase the risk for the infant born to a woman with epilepsy. Sixth and relatedly, fetal effects can vary depending on whether one or multiple AEDs are prescribed . Delineation of the teratogenic potential of each drug or drug combination is very difficult based on current evidence, especially for newer drugs. However, clear evidence exists to show that polydrug therapy is associated with a higher risk of congenital malformations than monotherapy, especially if polydrug therapy includes valproate. Additional evidence is needed to characterize the impacts of different polytherapy combinations, to assist the clinical management of women (and their infants) when monotherapy is ineffective for seizure control. Finally, there is increasing evidence to suggest that maternal genetic factors relating to the cerebral disorder underlying the epilepsy or other heritable conditions in the family increase the risk of congenital malformations. For example, a parental history of major congenital malformations increases the fetal risk more than fourfold, and the birth of a previous child with malformations when on the same AED increases the risk 17-fold.

With respect to individual AEDs, while the data are not perfectly consistent, several conclusions appear justified regarding the malformation and neurobehavioral risks associated with AED exposure during pregnancy. The most common major malformations reported in recent large series are neural tube defects, cardiac anomalies, oral clefts, hypospadias and other genitourinary anomalies, gastrointestinal anomalies, and skeletal deficits ( Table 38.9 ). The incidence of these malformations ranges from 2% to 10%, depending on the AED used, in contrast to a malformation rate of 1% to 3% in the general population.

Neonatal Effects

In relation to the neonatal effects of AEDs, neonatal withdrawal has been described. This has been reported most extensively with phenobarbital. Onset generally occurs around 7 days of age, which is understandable in view of the slow elimination of phenobarbital in the immediate neonatal period due to its half-life of several days. The major features consist primarily of CNS phenomena, including jitteriness, overactivity, disturbed sleeping, excessive crying, hyperreflexia, and disturbed sucking. Overt gastrointestinal phenomena, such as diarrhea, may occur but usually are not prominent. The symptoms may appear after the infant is discharged from the hospital, and the infant’s irritability may be mistaken for colic or some other extraneural cause. Symptoms often worsen over several weeks and persist for several months .

Neonatal withdrawal from other AEDs has been less extensively described. Recent reports describe withdrawal symptoms after prolonged in utero exposure to gabapentin. Symptoms developed within 24 hours of life, including sneezing, irritability, jitteriness, and loose stools. Given the increasing utilization of other novel AEDs, withdrawal should remain in the differential diagnosis for exposed infants presenting with similar symptomatology.

An additional potential complication for the newborn infant associated with maternal AED treatment using a select group of AEDs is neonatal hemorrhage . The drugs incriminated are hepatic enzyme inducers and have included hydantoins , barbiturates , primidone (which is metabolized to phenobarbital in vivo), and carbamazepine . In one series of 111 infants born to epileptic women treated with phenytoin or phenobarbital, 8 exhibited severe bleeding. In this syndrome, the infant developed hemorrhage shortly after birth. This early onset is unlike hemorrhagic disease of the newborn secondary to vitamin K deficiency. Sites of hemorrhage are, in order of decreasing frequency, skin, liver, gastrointestinal tract, intracranial sites, and thorax. Intracranial hemorrhage has been reported in 20% to 25% of the cases with any bleeding. The course may be fulminating, and approximately 40% of reported infants have died. Clotting studies demonstrate diminution of the vitamin K–dependent clotting factors (factors 2, 7, 9, and 10), with prolongation of either the prothrombin time or partial thromboplastin time, or both. Vitamin K levels in cord blood are depressed, as evidenced by the presence of PIVKA-II (a protein, induced by vitamin K absence, of factor II), an incompletely carboxylated, functionally defective form of factor II (prothrombin). Indeed, in one study of 24 infants born to mothers on antiepileptic therapy during pregnancy, 13 (54%) had detectable levels of PIVKA-II, and of the four infants exposed to valproate (none of whom had detectable levels of PIVKA-II), the incidence was 65%. Moreover, vitamin K levels also were depressed in the infants. These findings suggested increased degradation of vitamin K with impaired action of vitamin K on hepatic production of prothrombin. The latter occurs because of insufficient carboxylation of the glutamic acid residues of prothrombin, a posttranslational event that requires vitamin K. The pathogenetic mechanism by which some AEDs lead to vitamin K deficiency may relate primarily to increased degradation of vitamin K by fetal hepatic microsomal mixed-function oxidase enzymes, known to be inducible by phenytoin, phenobarbital, primidone, and carbamazepine. The similarity in structure of these so-called enzyme-inducible AEDs is shown in Fig. 38.8 . Nevertheless, in recent years this disorder appears to have become rare. Three reports ( n totals = 105, 204, and 662) have shown no increased incidence of hemorrhagic complications among women treated with various combinations of phenobarbital, phenytoin, primidone, and carbamazepine during pregnancy. The women were not treated with vitamin K during pregnancy, although the infants did receive vitamin K at birth. Whether the decline in incidence in this disorder relates to the diminishing use of polytherapy or increasing use of carbamazepine or related factors is unclear.

Chemical structure of important anticonvulsant drugs.

Note the structural similarities of phenobarbital, phenytoin, and carbamazepine; each is composed of a basic five- or six-membered ring and a side chain substitution containing phenol groups. Valproic acid differs markedly from the other anticonvulsant drugs and is essentially a short-chain fatty acid.

Cognition and Executive Functioning

Prospective studies have consistently documented the negative effects of valproate on cognitive development, with an 8- to 9-point IQ decrement. Of note, significant effects appear to relate largely to high-dose valproate exposure (>800 mg daily). The NEAD study also found that higher doses of valproate exposure during pregnancy were associated with poorer memory and executive function in offspring at age 6. In addition, evidence consistently suggests that monotherapy may be less harmful than polytherapy, although historically polytherapy has generally included valproate (see Table 38.10 ). Findings for other AEDs or combinations are mixed and insufficient.

Language Development

There is some suggestion that, based on IQ, verbal abilities may be more impaired than nonverbal abilities in children born to mothers treated with valproate. Of note, lower doses of valproate appear to be associated with less severe effects, although performance is still negatively affected compared with controls. Carbamazepine appears to exert similar negative effects as lower doses of valproate. In contrast, studies suggest that language abilities are comparable to control children for other AEDs, including phenytoin, lamotrigine, topiramate, gabapentin, and levetiracetam. However, these conclusions are based on studies with small sample sizes and unstandardized language measures.

Behavior and Regulatory Problems

A small series of studies suggest that children exposed prenatally to valproate are at increased risk of behavior problems based on screening measures, such as the Vineland Adaptive Behavior Scales and Strengths and Difficulties Questionnaire. Virtually nothing is known about specific risks for mental health problems, such as ADHD and anxiety disorders. One exception is a population-based Danish study that has linked valproate to an increased risk of autism spectrum disorder (ASD).

Motor and Visual Function

There is little evidence to suggest that prenatal AED exposure is associated with later motor problems, at least on global measures of psychomotor development. The one exception is a small study of 56 AED-exposed and 77 nonexposed newborns which found that exposed infants had lower limb and axial tone, and were less irritable than nonexposed infants. However, little evidence exists to suggest that prenatal AED exposure is associated with later motor problems, at least on global measures of psychomotor development.

Fetal Blood Flow

AEDs have been found to have direct cardiotoxic effects in fetal experimental models. Phenytoin and phenobarbital, specifically, inhibit the potassium channel IKr at clinically relevant concentrations. Via this mechanism, phenytoin induces bradycardia and transient ventricular arrhythmias in experimental models. These fetal cardiac effects may result in episodic hypoxia, leading to hemorrhage and necrosis of developing tissues. Preliminary data suggest lamotrigine has similar potential. The relevance of this adverse effect for other AEDs requires further investigation.

Fetal Malnutrition

There is also a suggestion that nutritional deficiencies may potentially modulate the teratogenic effects of AEDs. As discussed previously, hepatic enzyme inducers, including phenytoin, phenobarbital, and carbamazepine, have been implicated as a risk factor for neonatal hemorrhage. This effect may relate to increased degradation of vitamin K by fetal hepatic microsomal mixed-function oxidases inducible by these agents.

Women with epilepsy and abnormal pregnancy outcomes also have significantly lower blood folate concentrations. Valproic acid functions as a noncompetitive inhibitor of cellular folate receptors and increases methylenetetrahydrofolate reductase activity, which is a crucial determinant of folate utilization in the methyl cycle. Phenobarbital and carbamazepine decrease the expression of reduced folate carrier (RFC), potentially reducing folate uptake and transfer to the fetus. Newer AEDs also alter folate uptake and transfer. Specifically levetiracetam and lamotrigine decrease placental expression of RCF and folate receptor α, respectively. In rodent models of fetal valproic acid or hydantoin exposure, treatment with folate reduces the risk of congenital malformation. Unfortunately, studies in women with epilepsy do not suggest any protective benefits of folic acid supplementation before conception. In this context, preconceptional folic acid supplementation is recommended in a similar fashion as it is for all women of childbearing age for the prevention of neural tube defects (see Chapter 1 ). The role of higher doses in women with epilepsy has yet to be determined.

Molecular Effects

Several AEDs induce widespread neuronal apoptosis in a manner similar to fetal alcohol exposure. As described in detail earlier, the mechanism of alcohol-induced apoptotic neurodegeneration is likely mediated through NMDA receptor blockade and hyperactivation of GABA _A receptors . Similar neuronal apoptosis has been described in newborn rodents after exposure to phenytoin, phenobarbital, diazepam, and valproic acid, all of which directly activate GABA _A receptors or increase levels of GABA through sodium channel blockade or inhibition of catabolism. Of concern, recent in vitro data suggest similar potential with oxcarbazepine. Interestingly, although carbamazepine appears to exert antiepileptic activity through a similar mechanism, apoptosis is only augmented in rat pups exposed at postnatal day 7 at concentrations that far exceed those used in clinical practice. Notably, AEDs with a less clear impact on GABA, including levetiracetam and lamotrigine, appear to induce little to no apoptosis in experimental models. However, with the exception of levetiracetam, these safer agents do potentiate cell death when used in combination with proapoptotic AEDs.

An additional mechanism of teratogenesis may be the generation of reactive oxygen species (ROS). Experimental models demonstrate the ability of fetal peroxidases to rapidly bioactivate hydantoin derivatives (phenytoin and carbamazepine) to free radical intermediates that initiate the formation of ROS. Valproic acid also increases the formation of ROS in embryos, although the mechanism has not been clearly described. In vitro models demonstrate ROS generation in astrocytes after exposure to gabapentin, oxcarbazepine, and topiramate, but minimal effects from levetiracetam and lamotrigine. ROS generation interferes with development by direct oxidative damage to cellular macromolecules, including DNA and RNA. In addition, ROS generation leads to dysregulation of signal transduction, thereby triggering apoptotic or necrotic cell death.

Later in gestation, the fetus develops an additional ability to generate toxic metabolites via the cytochrome P-450 enzyme system ( Fig. 38.9 ). Several AEDs, including phenytoin, phenobarbital, carbamazepine, lamotrigine, and valproic acid, undergo metabolism in multiple phases. The first phase is generally a two-step hydroxylation catalyzed by cytochrome P-450. The first step produces a highly reactive epoxide that can bind nucleic acids and impair developmental processes. Multiple systems exist to detoxify harmful epoxides, including epoxide hydrolase, resulting in a hydroxylated molecule. Glucuronidation of these molecules in a later phase produces a highly water-soluble metabolite suitable for urinary excretion. Increased exposure to epoxides in experimental models increases the risk of teratogenesis. For example, exogenous inhibition of microsomal epoxide hydrolase produces a higher incidence of fetal demise and anatomical abnormalities after phenytoin exposure in mice. Importantly, several AEDs (carbamazepine, oxcarbazepine, phenobarbital, phenytoin, valproic acid) induce the cytochrome P-450 system, increasing the generation of reactive epoxides, and valproic acid specifically inhibits microsomal epoxide hydrolase (see Fig. 38.9 ). Many investigators have proposed these interactions as a potential explanation for the increased incidence of teratogenesis observed with polytherapy compared with monotherapy. Of note, inhibitors of the cytochrome P-450 system decrease the incidence of teratogenesis in experimental models. Importantly, this modulatory effect of polytherapy does not occur when using agents with no impact on the cytochrome P-450 system.

Potential mechanism of teratogenic effect of major anticonvulsant drugs.

Generation of epoxide derivative is catalyzed by a mono-oxygenase that is induced by the drugs themselves (i.e., autoinduction). The potentially teratogenic epoxide derivative must be degraded by a hydrolase to the apparently harmless dihydrodiol derivative. Impaired activity of the hydrolase, because of either genetic variability or exogenous factors (e.g., valproic acid), could cause the epoxide derivative to accumulate and exert teratogenic effects.

However, recent evidence suggests apoptosis may not be the fundamental mechanism of neurodevelopmental alterations from AEDs. These experimental models in mice highlight the importance of neocortical dysgenesis through alteration of proliferation and differentiation of neural progenitor cells. Specifically, valproic acid inhibits histone deacetylase, alters the expression of G1-phase regulatory proteins, inhibits neural progenitor cells from exiting the cell cycle during neocortical histogenesis, and increases the production of projection neurons in superficial neocortical layers. These findings highlight our incomplete understanding of the fundamental mechanisms underpinning well-described syndromes associated with valproic acid and other AEDs.

Genetic Factors

Genetic factors appear to mediate the incidence and severity of anomalies after prenatal AED exposure. Syndromic features are dramatically more likely to recur in children born to mothers with a previously affected offspring compared with mothers whose previous offspring showed no observable effects. Maternal metabolic variability may explain this observation to some extent. Cytochrome P-450 polymorphisms are common and result in more rapid production of toxic epoxides in some individuals. In addition, epoxide hydrolase exhibits polymorphisms that result in large variability in the capacity to detoxify metabolic intermediates. Observational studies of pregnant females taking phenytoin have found that lower epoxide hydrolase activity correlates with fetal hydantoin syndrome in offspring ( Fig. 38.10 ). However, the fact that these reactive hydrophilic metabolites most likely do not cross the placenta raises questions regarding these associations. In this case, maternal genetics may represent a proxy for fetal genetics, which clearly play a role in the risk of adverse outcomes associated with prenatal AED exposure. Children with major congenital anomalies after phenytoin exposure are also characterized by abnormal metabolite detoxification in experimental tests conducted in vitro after birth. Additional fetal genetic factors may include a sensitivity to the cardiac effects of AED exposure.

Epoxide hydrolase activity in amniocyte samples from 19 prospectively monitored fetuses whose mothers were administered phenytoin during the pregnancy.

The samples from the four fetuses subsequently noted postnatally to exhibit the clinical features of fetal hydantoin syndrome are shown in the front row of bars, and the 15 samples from fetuses subsequently confirmed not to have the characteristic features of the syndrome are shown in the rear row of bars. Note the markedly lower epoxide hydrolase activities in the fetal samples from infants who exhibited the clinical features of fetal hydantoin syndrome.

(From Buehler BA, Delimont D, Van Waes M, Finnell RH. Prenatal prediction of risk of the fetal hydantoin syndrome. N Engl J Med . 1990;322:1567–1572.)

Maternal-Fetal Effects

Cocaine use during pregnancy has deleterious effects on both the mother and the fetus ( Table 38.13 ). The specific features and the magnitude of these effects are relatively consistent across studies. A recent meta-analysis of 31 studies found that cocaine use during pregnancy was associated with significantly higher odds of preterm birth (OR = 3.4), low birth weight (OR = 3.7), and being born small for gestational age (OR = 3.2). On average, cocaine-exposed infants were born 1.5 weeks earlier and were almost 500 g lighter than nonexposed infants. Findings are mixed as to whether these infants catch up in growth, with one study finding no differences in the height and weight of exposed children between the ages of 1 and 6 years, and another reporting that they were in fact heavier at age 13 months, potentially indicating some postnatal compensation for intrauterine growth deficiencies.

Similarly, a recent meta-analysis of 10 studies found that methamphetamine exposure during pregnancy was associated with increased risks of preterm birth (OR = 4.1), low birth weight (OR = 4.0), and being born small for gestational age (OR = 5.8; see Table 38.13 ). Only three studies, two of which were from the same cohort, examined the extent to which these risks persisted after other confounding factors were taken into account. Findings from these studies showed that while risks were lower (OR range = 1.3 to 2.5), they remained statistically significant and consistent with unadjusted findings.

Neonatal Effects

The neonatal features of infants exposed to cocaine in utero are not characterized by a distinct constellation of craniofacial or other anomalies, but neurobehavioral features are common. The most consistent findings from both experimental and human infant studies indicate neurobehavioral and motor abnormalities. These consist principally of poorer state regulation, impaired arousal and altered sleep-wake states, poorer visual and auditory attentiveness, and increased periods of excitable and agitated behavior. Additional motor features include extensor and flexor hypertonicity, hyperreflexia, coarse tremor, and increased motor activity. Neither set of problems generally requires therapy, but rather appears to improve with time, especially the motor features. These adverse effects are even more evident when the fetus is exposed to other drugs of abuse, such as opioids, in addition to cocaine. The likelihood that these neurobehavioral characteristics are related to a direct effect of cocaine is supported by the demonstration of a dose-response relationship between the extent of maternal cocaine use during pregnancy and the severity of neonatal neurobehavioral and motor outcomes shortly after birth.

Methamphetamine exposed newborns are also characterized by a number of neurobehavioral features similar to those seen in cocaine-exposed infants. Neonatal intensive care unit (NICU) admission is often required. However, similar to cocaine, maternal methamphetamine use during pregnancy is generally not associated with a withdrawal syndrome that requires pharmacological treatment. Common features of methamphetamine withdrawal include a weaker and less coordinated suck, disorganized-state behavior, poorer quality of movement, and increased physiological/CNS stress. A dose-response relationship has been demonstrated for arousal and excitability. Severe symptoms requiring hospitalization resolve rapidly, with full resolution generally occurring by age 1 month.

An additional complication of prenatal cocaine exposure is sudden infant death syndrome (SIDS) , with this risk being approximately three- to sevenfold higher than for healthy unexposed infants. Both clinical and laboratory data suggest that the regulation of respiration and arousal are impaired in cocaine-exposed infants, and that such impairments could presage the occurrence of SIDS. Newborns exposed in utero to cocaine also have abnormalities of EEG and brain stem auditory evoked responses, which generally disappear after 1 to 6 months. However, more sophisticated EEG studies, based on quantified EEG, have shown persistence of abnormalities after 12 months. Limited experimental data suggest that the risk of SIDS may not be affected by methamphetamine exposure, although this conclusion requires further investigation.

Destructive Effects on the Central Nervous System

Cocaine appears to have both destructive effects (i.e., associated with the histopathological hallmarks of a destructive process, such as dissolution of tissue with a reactive cellular response) and development effects (i.e., teratogenic effects). It may be difficult to distinguish developmental from destructive effects in the fetus, because a destructive process that occurs in a rapidly developing tissue may not only have direct effects on already developed structure, but may also perturb the course of future developmental events. In addition, the reactive cellular response to tissue destruction may be less vigorous early in development, and thus evidence for a primarily destructive event may be absent or barely detectable morphologically. Indeed, distinction of the destructive effects from the teratogenic effects is difficult, in part because they may coexist in the same infant.

Cerebral Infarction and Intracranial Hemorrhage.

Cocaine and likely amphetamines are distinctive among the drugs that produce teratogenic effects on the CNS in their capacity to lead also to destructive neural effects . The first clearly documented example of a destructive lesion in the brain was the demonstration of infarction in the distribution of the middle cerebral artery in a newborn infant whose mother used cocaine by nasal insufflation in large doses during the 3 days before delivery. The presence of hemiparesis at birth and the computed tomography (CT) appearance of the lesion indicated that the infarct occurred shortly before birth, presumably during the period of cocaine exposure. Subsequently, numerous infants exposed to cocaine or methamphetamine in utero, with cerebral infarction in the distribution of major cerebral vessels, usually the middle cerebral artery, have been identified. ^a

a References .

In one series, 6% of cocaine-exposed infants exhibited cerebral infarction. Although all such lesions appear to have been prenatal in origin, the timing of the infarctions based on radiographical and clinical criteria has varied from hours to months before delivery. Thus, in the four affected newborns in the series of Dominguez and colleagues, three had well-established porencephaly, compatible with an event occurring weeks or months previously, and one had findings (edema) compatible with an acute event. A similar prenatal origin was apparent in an infant with hydranencephaly, related to destruction of the cerebral hemispheres in the distribution of the middle and anterior cerebral arteries, noted at birth after a pregnancy marked by cocaine exposure ( Fig. 38.11 ). However, it must be recognized that the reported cases were selected and were not identified in a prospective manner. In a prospective study of 717 cocaine-exposed infants, cranial ultrasonography, although not the optimal imaging modality, did not identify cerebral infarction. Smaller areas of apparent infarction have been described in basal ganglia, periventricular white matter, and brain stem by some investigators. However, it is unclear that such lesions are more common in cocaine-exposed infants versus appropriately chosen control infants.

Hydranencephaly in an infant exposed to cocaine in utero.

This computed tomography scan shows evidence of severe destruction of cerebral hemispheres in the distribution of the middle and anterior cerebral arteries. The appearance is consistent with severe ischemic insult in utero.

(Courtesy Dr. Elke Roland.)

Intracranial hemorrhage has been described in approximately 10% to 20% of primarily full-term cocaine or methamphetamine-exposed infants in one series, and in approximately 30% to 70% of very low-birth-weight, cocaine-exposed infants in three other series. Lesions appearing in more mature infants are not of major clinical importance. As previously noted, intracranial hemorrhage was not identified disproportionately in a prospective, controlled observational study of very-low-birth-weight infants. In summary, the available data derived from studies of cocaine or methamphetamine-exposed newborns by brain imaging lead to the tentative conclusion that both ischemic and hemorrhagic lesions may occur. However, available data indicate that the infarctions related directly to cocaine are likely exceedingly rare, and the hemorrhagic lesions are likely related largely to the degree of prematurity.

Teratogenic Effects on the Central Nervous System

The major abnormalities of CNS development reported after intrauterine cocaine exposure and the developmental event apparently affected (see Chapter 2 , Chapter 5 , Chapter 6 ) are summarized in Table 38.14 ).

Cognition and Executive Functioning

Most studies using measures of global cognitive functioning and IQ (e.g., Bayley Scales, Stanford Binet Test of Intelligence) have found that cocaine-exposed children generally perform in the low average score range . However, these cognitive effects appear to reflect other adverse exposures during pregnancy and/or in the postnatal environment, rather than the direct effects of cocaine. Similarly, children prenatally exposed to methamphetamine do not appear to be subject to cognitive delay or IQ deficits after covariate adjustment.

In contrast, when specific neuropsychological abilities have been examined in more recent studies, these suggest adverse effects on a range of abilities, including executive functioning, inhibitory control, working memory, and cognitive flexibility. ^a

a References .

Language Development

Global language delays are more common among children and adolescents exposed in utero to cocaine, with problems spanning both comprehension and receptive abilities . Importantly, these effects persist after covariate adjustment. Specific problems with syntax and phonology have also been reported. In contrast, no significant between-group differences were found between exposed and nonexposed 3-year-old children on measures of receptive or expressive language development after methamphetamine exposure. However, these children are at increased risk of educational difficulties in mathematics and language at a later age, although the extent to which these learning difficulties reflect direct drug effects or associated family risk exposures remains unclear.

Behavior and Regulatory Problems

Consistent with findings showing impacts on limbic and hypothalamic systems, cocaine exposure is associated with persistent emotional and behavioral regulatory problems spanning from infancy through adolescence. These difficulties manifest in a number of ways, including attention problems, aggression, and an increased risk of externalizing and internalizing behavior problems, including ADHD, conduct problems, anxiety problems, and depression. During adolescence, risk-taking and delinquent behaviors are common, such as fighting, running away from home, purposefully breaking or damaging things, early sexual behavior, and substance abuse. These latter problems are independent of other perinatal exposures and have been shown to be exacerbated and, in some cases, mediated by adverse family circumstances, including parental ongoing drug use, family violence, poor parental supervision, and caregiver mental health problems.

Although limited, studies of amphetamine exposure suggest similar long-term challenges. Prenatal methamphetamine exposure places children at increased risk of anxious/depressive problems and emotional reactivity during the preschool period. By school age, externalizing behaviors (aggressive behavior, rule-breaking) and ADHD symptoms are also more common, suggesting the likelihood of higher rates of ADHD later in life. Follow-up studies of methamphetamine-exposed adolescents are needed.

Motor and Visual Function

Findings are mixed with respect to motor outcomes of children exposed to cocaine during pregnancy, with studies largely confined to infants and young children. Some studies find no enduring effects of cocaine, whereas others report subtle impairments in fine motor and visual perceptual domains, especially for boys. As with other outcomes, these developmental difficulties frequently occur in the context of family psychosocial disadvantage, which predicts later risk more strongly than prenatal cocaine exposure.

Beyond the neonatal period, little is known about the potential adverse motor effects of prenatal amphetamine exposure. Poor fine motor and, in particular, grasping skills were noted at age 3 in a cohort exposed to methamphetamine. To date, studies in early childhood and beyond have not emerged.

Fetal Blood Flow

Cocaine and amphetamines act as stimulants by blocking the reuptake of dopamine, serotonin, and norepinephrine, leading to both euphoric and profound cardiovascular effects. Tachycardia and vasoconstriction in the mother, including direct vasoconstriction of the umbilical vein, impair placental blood flow , disrupting oxygen and nutrient transfer to the fetus. These acute alterations in fetal blood flow may contribute to the destructive effects of cocaine in the neonatal CNS. Increases in fetal cerebrovascular resistance have also been documented following cocaine administration. Although global cerebral blood flow is not disturbed, the state of development of cerebral vessels in the human fetus may explain the distribution of ischemic lesions observed in human studies. The extraparenchymal, leptomeningeal arteries begin to develop a distinct muscularis in the second trimester and have a well-developed muscularis in the third trimester of gestation, the time of occurrence of the cerebral infarcts reported in the cocaine-exposed infants. Thus the middle cerebral artery, the vessel affected in most cocaine-induced strokes, presumably does not develop the capacity to undergo spasm until the third trimester. Intracranial hemorrhage could also occur in the immediate neonatal period secondary to elevation in blood pressure and cerebral blood flow velocity that have been documented on the first and second postnatal day in infants exposed to cocaine in utero. Elevated levels of catecholamines in the presence of cocaine also increase uterine contractility, likely resulting in the increased rate of spontaneous abortion, preterm birth, and placental abruption associated with antepartum cocaine use.

Fetal Nutrition

Cocaine and amphetamines have important impacts on fetal nutrition , arising from both maternal malnutrition as well as direct effects. Cocaine and amphetamines are both associated with low body mass index in adults, potentially due in part to their anorexic effects. Cocaine also inhibits the placental transport of specific amino acids, including arginine, phenylalanine, and valine. Of interest, these effects appear to be exacerbated by concurrent exposures, including nicotine. In addition, catecholamines increase metabolism of nutrients, further contributing to fetal depletion.

Molecular Effects

Cocaine administration to neonatal rats results in the inhibition of DNA synthesis and alterations of membrane lipids in all brain regions, consistent with the direct effects shown in nonneural tissues. Deleterious effects on neuronal differentiation in rodent models have also been observed. The disturbances in neuronal number and differentiation have also been observed in primate models.

The mechanism of the effects of cocaine on neuronal differentiation and on development of crucial neuronal pathways in the cerebrum, diencephalon, and brain stem (subsequent to neuronal proliferation) could relate to the fundamental mechanisms of the effects of cocaine on neurotransmitters. The neuronal pathways involved either use as neurotransmitters several monoamines (e.g., norepinephrine, dopamine, serotonin) or are the targets of neurons that use monoamines as neurotransmitters. The derangements of homeostasis of these neurotransmitters by cocaine are, of course, the principal mechanisms of action of the drug in the adult brain. The specific nature of the derangements in the fetus requires further definition but appears similar. Although increased levels of these neurotransmitters might be expected, at least initially, chronic exposure to cocaine could lead to their depletion, as has been shown in experimental models and in adult humans. The findings of reduction in striatal dopamine in neonatal rabbits exposed to cocaine in utero and diminished cerebrospinal fluid levels of homovanillic acid, the principal metabolite of dopamine, in cocaine-exposed newborns supports this possibility. These neurotransmitters appear very early in brain development and play important regulatory roles in development of neuronal circuitry. Indeed, because these monoaminergic systems, which originate especially in the brain stem, have widely distributed contacts in the basal ganglia, cerebral cortex, hypothalamus, and elsewhere, disturbances in their function during development could have very far-reaching effects. For example, neurons containing norepinephrine in the locus ceruleus (a pigmented nucleus in the pons) give rise to ascending monosynaptic pathways that are distributed widely in the cerebral cortex and diencephalon. These pathways are considered to have primarily activating influences. A prominent dopaminergic system originates in the substantia nigra of the midbrain and terminates in the striatum. These connections are important in the regulation of movement and tone. Other dopaminergic systems terminate in limbic and cortical structures and are crucial not only for the reinforcing actions of drugs like cocaine but importantly also for motivational-attentional functions. Prominent serotoninergic systems originate in the raphe nuclei of the brain stem, project widely, and are crucial for regulation of sleep and level of alertness. However, because disturbances of the development of neuronal circuitry do not leave a readily identifiable, morphological stamp, and require highly sensitive immunochemical and neuropharmacological techniques for detection, identification of such neural abnormalities in humans is very difficult. However, impairment of subsequent cerebral cortical development by prenatal cocaine exposure during the period in which monoamines influence neural development has been shown in the primate as well as rodent cerebral cortex. Careful study of cerebral cortical development in the rat has shown in the adult cortex abnormalities of neuronal lamination and dendritic morphology after prenatal exposure to cocaine (see Fig. 38.12 ).

Genetic and Epigenetic Alterations

Whether the effects on neuronal proliferation and differentiation relate to perturbation of the action of critical genes , including immediate early genes, is an intriguing possibility. Cocaine exposure downregulates the placental norepinephrine transporter (NET), responsible for sodium-chloride-dependent reuptake of extracellular norepinephrine. Reduced placental NET expression may lead to increased circulating norepinephrine, downregulation of the steroid metabolic enzyme 11β-HSD-2, and fetal hypercortisolism, resulting in altered activity of the hypothalamic-pituitary-adrenal (HPA) axis. Altered function of the HPA axis has been implicated as a potential contributor to long-term behavioral and emotional disorders observed in children exposed prenatally to cocaine. These changes are associated with hypermethylation of genomic DNA, most profoundly in promoter regions. In adults, chronic cocaine exposure permanently alters brain-derived neurotrophic factor (BDNF) and Cdk5 promoters. Alteration of these and other promoters has the potential to produce profound alterations in development throughout life. Of interest, genetic modifications may not be confined to the maternal-fetal dyad, as evidenced by behavioral alterations in offspring after paternal cocaine exposure.

Maternal-Fetal Effects

Among infants born to heroin-dependent women, the incidence of low birth weigh t (i.e., <2500 g) was approximately 40% to 50% among early studies. Relatedly, approximately one-third of exposed newborns will be small for gestational age (SGA), and approximately 40% will have a head circumference below the 10th percentile for gestational age ( Table 38.16 ). This occurrence of intrauterine growth restriction in heroin-exposed infants is reminiscent of experimental studies showing growth retardation in the progeny of female rats given morphine before but not after conception. However, it is likely that other factors, related to nutrition and infection, were present in the mothers before and during pregnancy. Unfortunately (and as highlighted earlier), many of the early heroin studies did not examine the relative contribution of these correlated lifestyle factors.

Infants born to methadone-maintained mothers have been shown to be at increased risk of being born early, and even when born close to term, to weigh less, be shorter, and have smaller head circumferences than infants born to non-drug-using mothers. Between 10% and 35% are low birth weight (i.e., <2500 g). Of these, two in five infants will be born SGA. Risks of SIDS, strabismus, hyaline membrane disease, vision abnormalities (particularly nystagmus), and congenital birth defects are also higher among infants born to mothers receiving methadone maintenance treatment compared with infants born to non-drug-dependent mothers. However, findings demonstrate that outcomes associated with maternal methadone use are generally better than for untreated heroin use during pregnancy. The reasons for the disturbances in intrauterine growth are not entirely clear, and the relative roles of intrauterine undernutrition, infection, and toxic effects of other drugs and exogenous materials remain to be elucidated. Although a direct effect of narcotic analgesics on cell number has been suggested by experimental studies (discussed in a later section), findings are mixed with respect to the analysis of human infants. Some studies are more supportive of an indirect effect related to impaired maternal nutrition and other factors, whereas other studies support the persistence of growth differences, even after covariate factors are taken into account.

Only a few studies have examined the neonatal effects of maternal buprenorphine treatment during pregnancy. Of these, some have found no differences in the risks of fetal death, preterm birth, low birth weight, and SGA/growth restriction, whereas others have reported a lower risk of preterm birth and higher birth weights for buprenorphine-exposed infants compared with methadone-exposed infants. A recent meta-analysis of 18 studies (three randomized controlled trials [RCTs] and 15 cohort) found that buprenorphine was associated with lower risk of preterm birth (RR = 0.40 to 0.67), greater birth weight (weighted mean difference = 265 to 277 g), and a larger head circumference (weighted mean difference = 0.68 to 0.90), with bigger differences observed in RCT studies. No differences were found for spontaneous fetal death, fetal congenital anomalies, and other growth measures.

Neonatal Abstinence Syndrome

Early studies suggested that the incidence of NAS in heroin-exposed infants was about 70%, varying from about 50% to 90%. ^a

a References .

b References .

Concerning the perinatal outcomes associated with maternal prescription opioid use, findings are now emerging but are exclusively confined to medical record review. For example, an analysis of more than 100,000 pregnant women who filled more than one prescription (28%) in Tennessee found that opioid-exposed infants with and without NAS (the latter implying high levels of exposure) were more likely to be born with low birth weight (no NAS 21.2%, NAS 11.8%) than nonexposed infants (9.9%).

The clinical presentation of NAS following opioid exposure consists of an array of signs and symptoms, including increased irritability, hypertonia, tremors, seizures, feeding intolerance, watery stools, emesis, and respiratory distress, with both male and female infants affected equally. The predominance of signs and symptoms relevant to the CNS and gastrointestinal tract relate to the particular concentration of opioid receptors in these regions (see later). These symptoms tend to emerge between 6 and 48 hours after birth as a result of the sudden discontinuation of exposure to the opioids being used or abused by the mother during pregnancy ( Fig. 38.15 ). The time of onset of the withdrawal syndrome in the newborn delivered to a heroin-addicted woman is usually quite early ( Table 38.17 ). Approximately 65% of infants present within the first 24 hours of life; an additional approximately 20% on the second day of life; and the remainder, or about 15%, on the third and fourth days. This contrasts with the time of onset of the withdrawal symptoms with methadone. Most infants exposed to methadone have the onset of overt symptoms on the second day of life ( Table 38.18 ), not the first day as with heroin. Median onset of therapy in a recent study was 35 hours. Ten percent to 15% of infants born to methadone-treated mothers have had onset of their neurological syndrome after day 3. This late onset is rare with infants passively addicted to heroin. The delay in onset correlates with the administration of the last dose of methadone before the time of delivery ( Fig. 38.16 ). Indeed, withdrawal symptoms in infants passively addicted to methadone may occur initially or may recur as late as 2 to 4 weeks or more after birth. In fact, an occasional infant has died before the nature of this later syndrome has been recognized.

Onset and duration of neonatal abstinence syndrome symptoms from heroin, methadone, and buprenorphine.

Note the delayed onset and prolonged duration of methadone withdrawal compared to heroin withdrawal. Interestingly, despite a long half-life, buprenorphine withdrawal symptoms appear early and resolve rapidly compared to methadone.

(Adapted from Binder T, Vavrinkova B. Prospective randomised comparative study of the effect of buprenorphine, methadone and heroin on the course of pregnancy, birthweight of newborns, early postpartum adaptation and course of the neonatal abstinence syndrome (NAS) in women followed up in the outpatient department. Neuro Endocrinol Lett . 2008;29:80–86.)

Passive addiction to methadone: relation of the time of the last maternal methadone dose (LMD) to the time of onset of neonatal withdrawal.

The interval between onset of withdrawal and time of last maternal dose did not vary markedly and approximated 48 hours.

(From Rosen TS, Pippenger CE. Pharmacologic observations on the neonatal withdrawal syndrome. J Pediatr . 1976;88:1044–1048.)

The reasons for the delayed onset and prolonged duration may relate to the pharmacokinetics of methadone elimination in the newborn. It is well known in adults that the withdrawal syndrome in methadone addiction develops more slowly (peak 6 days) and lasts longer than with heroin. Good evidence is available for avid tissue binding of methadone, with gradual and slow release on withdrawal. The critical point is that the clinician must be alert to this possibility of delayed onset in the infant of age 1 week or more who develops jitteriness, diarrhea, and other signs of withdrawal.

The likelihood of withdrawal symptoms in an infant exposed to heroin during pregnancy seems to relate primarily to five factors: (1) the amount of the maternal dose; (2) the length of the time from the last dose to birth; (3) the duration of maternal drug use/abuse; (4) the use of other drugs, particularly nicotine and SSRIs; and (5) the gestational age of the infant at birth. ^a

a References .

Similarly, the gestational age of the infant at birth influences the likelihood and severity of neonatal withdrawal after methadone ( Fig. 38.17 ). In contrast to heroin, maternal methadone dose does not appear to be a critical determinant of the likelihood of withdrawal, although cord blood levels are inversely related to the occurrence of a neonatal withdrawal syndrome requiring therapy. In fact, the likelihood of neonatal withdrawal features directly correlates with more rapid rates of decline of neonatal methadone levels. Moreover, the severity of symptoms also relates particularly to the rate of elimination ( Fig. 38.18 ). This finding is entirely consistent with an observation in adult patients (i.e., a rapid decrease in the drug level in blood causes more frequent and more severe symptoms).

Relationship between the severity of the neonatal withdrawal syndrome after passive addiction to methadone and the gestational age of the infant.

Scores are expressed as median values, with ranges in parentheses. Note the increase in severity with advancing gestational age.

(From Doberczak TM, Kandall SR, Wilets I. Neonatal opiate abstinence syndrome in term and preterm infants. J Pediatr . 1991;118:933–937.)

Passive addiction to methadone: relation of occurrence and severity of signs of withdrawal to rate of elimination of methadone.

Data derived from study of 31 infants born to mothers receiving methadone during pregnancy. The fastest declines in plasma methadone (M) concentrations correlated with occurrence of the most severe signs.

(From Rosen TS, Pippenger CE. Pharmacologic observations on the neonatal withdrawal syndrome. J Pediatr . 1976;88:1044–1048.)

The initial and dominant symptoms and signs of the withdrawal syndrome relate primarily to disturbance of the CNS ( Table 38.19 ). A virtually invariable feature is jitteriness. As discussed in Chapter 12 , the movements of jitteriness are characterized primarily by tremulousness; are exquisitely stimulus-sensitive, rhythmic, and usually of equal rate and amplitude; and can be induced to cease by gentle passive flexion of the limb. Unlike seizure, jitteriness is accompanied neither by abnormalities of gaze or extraocular movement, nor by the clonic jerking of limbs. The tremulous movements in infants passively addicted to heroin are usually quite dramatic and have a coarse, flapping quality. The babies are very irritable and frequently are extremely active and hypertonic. Sleep periods are markedly diminished, and the cry is often high pitched and shrill. Sucking is excessive, and often frantic sucking of the fists or fingers has been noted. These frequent clinical features (75% to 100% of symptomatic cases) are very helpful in making the clinical distinction of the jitteriness of drug withdrawal from that resulting from other important causes (e.g., hypoglycemia and hypoxia). Hypoglycemia can cause jitteriness in the first 24 to 48 hours of life, but with hypoglycemia the infant is almost always stuporous as well, quite unlike the hyperalert, hyperactive infant with withdrawal symptoms. Similarly, hypoxic-ischemic encephalopathy characteristically causes jitteriness in the first 24 to 48 hours of life, but such infants almost always have had the characteristic history of a perinatal hypoxic-ischemic insult, are significantly stuporous, and very frequently exhibit seizures. The clinical picture may be complicated by concomitant passive addiction and hypoxic encephalopathy; in several series, 20% to 40% of addicted infants exhibited signs of fetal distress, Apgar scores of six or less at 1 or 5 minutes, or both.

Next in frequency of occurrence are a series of symptoms relating especially to gastrointestinal disturbance (see Table 38.19 ). Poor feeding is the most prominent of these features (one that in fact may reflect more of a neurological than a gastrointestinal disturbance). Despite the initially excessive sucking described previously, the infant decreases its sucking rapidly with feeding. Poor coordination of sucking, swallowing, and respiration is very common. Regurgitation of feedings is also a relatively common feature. Diarrhea occurs in as many as 30% to 50% of the infants and can contribute to dehydration and electrolyte disturbances. These latter gastrointestinal phenomena tend to appear later (i.e., at days 4 to 6) than those related to the CNS.

Sneezing and tachypnea are less common disturbances (see Table 38.19 ). Still less common, but disturbing when they do occur, are fever and sweating . Fever must always raise the possibility of infection, and this possibility should be ruled out by appropriate diagnostic studies. Sweating, while uncommon, can be a helpful diagnostic sign because it is very unusual to see sweating in newborns, especially small newborns.

Seizures are a distinctly uncommon manifestation of neonatal withdrawal to opioids (see Table 38.19 ). The incidence has been approximately 1% to 2% of cases. It is in fact difficult to be convinced from the published data that any of the examples of seizure associated with withdrawal to heroin were not related to complicating factors, such as hypoxia-ischemia or metabolic disease, or were not examples of particularly marked jitteriness. This conclusion is compatible with the fact that seizures are not a feature of withdrawal to heroin in adult patients. Thus it should be emphasized that passive addiction to opioids should be low on the list of considerations when one is faced with a newborn with seizures, and even in the infant definitely passively addicted to heroin, seizure phenomena should raise the possibility of a serious complicating illness and provoke appropriate diagnostic studies.

Cognition and Executive Function

Studies examining the cognitive outcomes of children exposed to opioids during pregnancy report conflicting results. While some studies report no difference, others have found that opioid-exposed children perform less well on standardized cognitive measures. ^a

a References .

Behavior and Regulatory Problems

There is fairly consistent evidence, at least based on parent and teacher report, to show that children exposed to opioids during pregnancy have higher levels of attention, hyperactivity, and impulsivity problems than non-opioid-exposed children. There is also some suggestion that these problems may become more problematic with age. However, the proportion of opioid-exposed children who meet either DSM or ICD criteria for a range of potential psychiatric disorders, including anxiety and conduct disorder, is not known.

Motor and Visual Function

There is increasing evidence that infants born to opioid-using mothers may be at increased risk for a range of visual abnormalities, including nystagmus, strabismus, reduced visual acuity, delayed visual maturation (DVM), impaired voluntary eye movements, and absent binocular vision. A recent study of 81 methadone-exposed and 26 nonexposed comparison infants at age 6 months (retention 79% and 52%, respectively) found that opioid-exposed infants had a fivefold increased risk of failing standardized ophthalmologic assessment, with 40% (32/81) failing and a further 11% (9/81) deemed borderline. Visual abnormalities included horizontal nystagmus ( n = 9/32), strabismus ( n = 20/32), and reduced visual acuity (>0.90 logMAR) ( n = 18/32). The latter was associated with other visual abnormalities in 11 of these infants. Further electrophysiological assessment of visual function using visual evoked potentials (VEP) found that 70% of opioid-exposed infants exhibited abnormal VEP parameters (slower peak times, smaller amplitudes), suggesting that prenatal drug exposure may have altered the functioning of visual pathways and/or cerebral sources of VEPs (see Chapter 10 ). These adverse clinical and electrophysiological outcomes were independent of other prenatal drug exposures.

Pathogenesis of Neonatal Abstinence

Opioids act through G protein–coupled receptors µ, κ, and δ; µ-opioid receptors appear to predominate in the developing brain. Neurons in the locus coeruleus of the pons have a high density of noradrenergic µ-opioid receptors, and this nucleus appears to be the principal site that triggers opioid withdrawal. µ-Opioid receptor stimulation suppresses production of cyclic adenosine monophosphate (cAMP), decreasing the release of norepinephrine from neurons in the locus coeruleus. The persistent presence of an opioid stimulus eventually leads to the compensatory release of norepinephrine, possibly through upregulation and/or supersensitization of adenylyl cyclase isoforms (described as development of alternative pathways and disuse hypersensitivity in the previous edition of this text). Upregulation/supersensitization appears to rely on phosphorylation of the opioid receptor after persistent stimulation, followed by uncoupling from the G protein and receptor colocalization with β-arrestins, protein kinases (including activated Src), and adenylyl cyclases. In the absence of depressive opioids, hyperactive neurons in the locus coeruleus release excessive amounts of norepinephrine, resulting in hyperthermia, hypertension, tremors, and tachycardia.

Other sites of altered enzymatic activity include the ventral tegmental area of the midbrain, the medial habenula of the thalamus, the hypothalamus, and the dorsal raphe nucleus of the brain stem. The ventral tegmental area releases less dopamine during opioid withdrawal, resulting in hyperirritability and anxiety. The midbrain and medial habenula release increased amounts of acetylcholine, contributing to diarrhea, vomiting, and diaphoresis. The hypothalamus releases increased amounts of corticotrophin, leading to increased stress and hyperphagia. The dorsal raphe nucleus releases decreased amounts of serotonin during opioid withdrawal, leading to sleep disturbances. Decreased levels of brain-derived neurotrophic factor and corticotrophin releasing-factor appear to underlie the latter; however, the mechanisms of all of these alterations remain incompletely elucidated.

Pathogenesis of Potential Long-Term Consequences

The molecular effects of opioid exposure include apoptosis as well as altered patterns of myelination. In vitro studies demonstrate the apoptotic effect of even a single dose of opioid on neural progenitor cells, increasing levels of caspase-3 in these proliferating cells. Opioids disrupt myelination through multiple mechanisms. Opioids accelerate maturation of preoligodendrocytes in the developing brain, a phenomenon that appears to disrupt the synchronized sequence of events contributing to normal connectivity. In addition, opioids suppress neuronal migration to the cortical plate, and alter dendritic growth and branching patterns. The underlying mechanisms of these effects are not completely understood, although they likely reflect, at least to some degree, suppression of activity-dependent network formation as well as phosphorylation and uncoupling of the opioid receptor after persistent stimulation. Receptor colocalization and binding with β-arrestin activates signal transduction, recruiting kinases. Formation of a receptor/β-arrestin/extracellular signal-regulated kinase (Erk) aggregate inhibits the growth-promoting effects of Erk. Alternatively, receptor/β-arrestin may bind to c-Jun N-terminal kinase and apoptosis signal-regulating kinase, increasing the activity of these mediators of cell death. In addition, chronic opioid exposure results in lower levels of BDNF in the hippocampus. Decreased levels of BDNF precipitate decreased activation of tropomyosin receptor kinase (Trk) receptors, expressed widely in this brain region. Trk receptors are responsible for the release of Akt serine/threonine kinase, an essential inhibitor of the apoptotic caspase cascade. Cumulatively these, and likely other perturbations, result in the phenotypic neurodevelopmental manifestations of fetal opioid exposure.