Pharmacogenomics, the study of specific genetic variations and their effect on drug response, will likely give rise to many applications in maternal–fetal and neonatal medicine; yet, an understanding of these applications in the field of obstetrics and gynecology and neonatal pediatrics is not widespread. This review describes the underpinnings of the field of pharmacogenomics and summarizes the current pharmacogenomic inquiries in relation to maternal–fetal medicine—including studies on various fetal and neonatal genetic cytochrome P450 (CYP) enzyme variants and their role in drug toxicities (for example, codeine metabolism, sepsis and selective serotonin reuptake inhibitor (SSRI) toxicity). Potential future directions, including alternative drug classification, improvements in drug efficacy and non-invasive pharmacogenomic testing, will also be explored.

With the rise of increasingly personalized medicine as a backdrop, the field of pharmacogenomics—the study of specific genetic variations that affect drug metabolism, response and action—stands out as especially promising. Perhaps no other area within this field generates greater excitement and potential for altering the way we practice medicine than fetal and neonatal pharmacogenomics. Although only in its infancy (no pun intended), this area will not only shape the way we diagnose and treat diseases of expectant mothers and their infants, but will also improve our overall understanding of fetal and neonatal development. For example, by focusing our attention on variations in genetic code and gene expression between pre- and post-natal life, pharmacogenomics has the capacity to contribute greatly to our understanding of cellular, tissue and organ differentiation, physiological and pathological fetal development and variation in fetal and neonatal drug metabolism.

Fetal exposure to maternal medications may occur as a result of deliberate maternal drug therapy or inadvertent drug exposure. Since the discovery that linked thalidomide with birth defects, obstetricians and their expectant patients have had to weigh the balance between maternal drug therapy and potential fetal exposure, often with limited information about the effects of this exposure. Although some medications may be suspended during pregnancy or replaced with safer alternatives, others, such as antiepileptic and antithrombotic medications, are required to prevent serious maternal complications and may be difficult to discontinue without putting the mother at increased risks.¹

As we begin to understand the adverse fetal effects of specific medications, it is clear that not all fetuses exposed to a given medication will bear evidence of its associated malformation(s), and that these effects may be varied in both their presentation and timing. Therefore, the capacity to identify fetal risk in an increasingly sensitive manner would greatly benefit the specificity with which we will be able, as physicians, to recommend courses of pharmacotherapy during pregnancy and the postpartum period. In this review, we provide a summary of the progress that has already been made in the field of maternal–fetal and neonatal pharmacogenomics and explore a few of its potential future directions.

Pharmacogenomics is a field of a study in which multiple genes, none of which dominate, are explored in relation to complex drug action and response, frequently in the context of pharmacokinetics (drug disposition) and pharmacodynamics (drug action). Pharmacogenetics, on the other hand, describes situations in which a single drug reaction is accounted for by one or a few genes. Although pharmacogenomics has existed for decades, the field has recently experienced a significant upsurge, partly because of technological advances that make genotyping and sequencing increasingly time and cost effective. Both the popular and scientific media have followed these developments as an indication of our movement toward ‘personalized medicine’. Indeed, the potential to treat diseases in such a specific, targeted manner is a powerful goal in the world of medicine.

A full exploration of the field of pharmacogenomics is beyond the scope of this review, but a brief introduction will help to guide the next sections. In short, pharmacogenomics tries to correlate specific genetic variations, most frequently in the form of single-nucleotide polymorphisms, copy number variations, deletions and duplications, with an associated drug response. It is generally understood that these variations can result in altered protein structure and function and therefore lead to unique pharmacokinetic and/or pharmacodynamic profiles in individuals. For example, genes with certain single-nucleotide polymorphisms may have altered products, including enzymes, receptors, neurotransmitters, growth factors, transporters and other proteins. In this context, the cytochrome P450 (CYP) enzymes, which are found predominantly in hepatic tissues (but also in smaller concentrations in a variety of extrahepatic tissues) and are involved in drug bioactivation and metabolism, have a prominent role. CYP genes are followed by an Arabic numeral indicating the gene family, a capital letter indicating the subfamily, and another numeral for the individual gene. The Cytochrome P450 Nomenclature Committee currently designates genetic variants with an additional numeral punctuated by an asterisk (for example, CYP2J2^*2) and the corresponding protein variants with a numeral punctuated by a period (for example, CYP2J2.2).

When certain genetic variations can be linked with specific fetal outcomes or toxicity risk profiles, individual gene sequencing may provide a blueprint for effective and/or safe maternal pharmacotherapies on an individualized or population basis. The development and application of this field poses both feasibility concerns and ethical questions that have yet to be solved, but also offers the potential to wield greater control and specificity in medical care.^{2, 3}

Before we can begin to describe specific developments in maternal–fetal and neonatal pharmacogenomics, it is important to place fetal exposure to drugs in the context of normal fetal and placental development. The pre-implantation period, which is the 2-week period from fertilization to implantation, has often been called the ‘all or none’ period (Table 1). A large insult at this stage of embryogenesis will usually result in death of the conceptus.⁴ The next period, from the third week through the eighth week of gestation, is called the embryonic period and represents the stage when the main organ systems (including the central nervous system, cardiac and gastrointestinal systems) are established.⁵ Thus, the embryonic period is the most crucial in term of structural malformations. The period from the ninth week to birth is known as the fetal period and is characterized by tissue and organ maturation, as well as rapid growth of the body. Fetal drug exposure during this period may result in congenital malformations and/or growth abnormalities.

It is important to recognize that differences in gene expression occur as the embryo develops.¹ Thus, variations in gene expression may only be identified and relevant at specific critical points in the developmental continuum. This process continues well into the neonatal period and early childhood; thus, pharmacogenomic variations discovered in adults may not be relevant for neonates and children. A good example of this principle exists in the effects of fetal exposure to warfarin (coumadin). Exposure between the sixth and ninth week increases risk for warfarin embryopathy, which is characterized by nasal hypoplasia and stippled vertebral and femoral epiphyses. It is unlikely that these defects result from fetal hemorrhage because vitamin K clotting factors are not demonstrable in the embryo at this age. Instead, it is thought that the teratogenic effect occurs as a result of inhibition of post-translational carboxylation of coagulation proteins.⁶ However, fetal defects due to warfarin exposure in the second and third trimesters (including dorsal central nervous system dysplasia and developmental and mental retardation) likely result from hemorrhage leading to scarring.

Several layers of defense exist to protect the developing embryo from drug exposure and toxicity. The first of these is maternal drug metabolism and biotransformation. In addition to terminating pharmacological activity, biotransformation enhances drug elimination from the body, reducing the amount of parent compound available to cross the placenta.¹ Thus, maternal liver and/or renal disease must be considered when using potential toxic medications. For example, pre-eclampsia, a pregnancy-associated disorder characterized by generalized vasoconstriction, may lead to both maternal hepatic dysfunction and renal arteriolar vasoconstriction and thus decreased urinary excretion. As a result, increased circulating levels of medications may lead to greater maternal side effects and/or neonatal effects.

The second line of defense is the placenta. Substances that pass from maternal blood to fetal blood must traverse syncytio- and cytotrophoblast layers, intravillous stroma and the fetal capillary wall.⁴ Several variables are important in the effectiveness of the human placenta as an organ of transfer, including the concentration of substance in maternal plasma, the rate of maternal blood flow through the intervillous space, the area available for exchange, the amount of substance metabolized by the placenta during transfer, the specific binding or carrier proteins in the fetal or maternal circulation and the rate of fetal blood flow through the villus capillaries.⁴ For substances transferred by diffusion or active transport, the physical properties of the tissue barrier or capacity of the biochemical machinery of the placenta for effective active transfer, respectively, are important. In general, most substances with a molecular mass <500 Da diffuse readily through placental tissue, and lipophilic molecules tend to cross more readily.⁴

Besides acting as a mechanical barrier, and because of its origin in fetal trophectoderm, the placenta expresses various enzymes involved in drug metabolism, including CYP enzymes, superoxide dismutase, glutathione peroxidase, glutathione reductase and glutathione S-transferases.¹ Allelic variations in these enzymes have the potential to affect fetal exposure to exogenous compounds, and therefore individual susceptibility to adverse consequences in the fetus and newborn.¹ Examples of alleles with variable placental CYP gene expression include CYP2J2, CYP4B1, CYP2C19, CYP2C9, CYP2D6 and CYP3A7.^{7, 8} Upregulation of placental CYP1A1 and CYP2E1 in rats have also been shown in pathological conditions such as nicotine exposure.⁹ Figure 1 characterizes the potential effects of maternal, placental and fetal pharmacogenomics on antenatal drug exposure and toxicity.

Maternal–fetal pharmacogenomics: effects on antenatal drug metabolism and exposure.

Full size image

Approximately 3% of all infants are born with one or more major birth defect.^{10, 11} Although the risk of exposure to ‘high-risk’ teratogens, such as thalidomide or isotretinoin, is low, it is currently unclear what predisposes certain fetuses to the harmful effects of the more commonly used ‘moderate-risk’ medications, such as anticonvulsants and anticoagulants (Table 2).¹² In addition, 7% of children in the United States have developmental disorders by age 1, 12–14% by the time they enter school and 17% before age 18; the role that in utero drug exposure may have in developmental disorders continues to be studied.¹³

It is important to note that the same enzymes that are currently hypothesized to be involved in altered pregnancy outcomes (for example, androgens, cholesterol, eicosanoids, estrogens, progestins, retinoic acid, thyroxine, vitamin D and so on) also have critical roles in the synthesis and catabolism of endogenous compounds that are important for fetal growth and development.¹² Thus, medications administered with therapeutic intent have the potential to disrupt normal cellular function through competition with endogenous substrates and ligands.

When evaluating the role of fetal pharmacogenomics in adverse birth defects, it is important to analyze a variety of fetal tissues for variations in gene expression (although CYP genes are expressed mainly in hepatic tissues, extrahepatic expression occurs in kidney, lungs and heart, among many other tissues); however, to date, this has not been routinely performed. For example, Bieche et al.,⁷ who examined CYP 1, 2 and 3 mRNA from different human tissues, including fetal liver, showed that contrary to adult liver—in which several CYP enzymes are expressed—CYP3A7 predominated in fetal liver and was nearly exclusive to this tissue. However, no other fetal tissues were assayed. In methodological contrast, a study by Gaedigk et al.¹⁴ compared CYP2J2 mRNA expression in a variety of fetal tissues and found fetal liver and heart levels comparable. Thus, it may be important not to rely solely on fetal liver assays when analyzing CYP gene expression patterns and activity.

It is also important to distinguish fetal CYP gene expression from gene and protein expression in neonates and adults. In their study of CYP2J2, an enzyme that metabolizes arachidonic acid, Gaedigk et al.¹⁴ found that CYP2J2 mRNA is expressed in human fetal liver as early as in 11-week gestation and that there were differences in pre- and postnatal protein expression patterns. Although CYP2J2 gene expression was similar in their pre- and postnatal samples, the postnatal samples showed that CYP2J2 antibody recognized two or three distinct bands. Differential protein patterns between pre- and postnatal liver were also observed with a second antibody (CYP2J2pep3). The researchers explored this finding by genotyping for known (and unknown) CYP2J2 variants and also by detecting aberrant mRNA splice variants without success. Their group also showed that inter-individual variability is relatively low.¹⁴

Another example of fetal pharmacogenomics exists in a recent analysis of acetaminophen metabolism in the fetus and the theoretical risk of gastrischisis, a congenital defect of the abdominal wall.¹⁵ Acetaminophen is eliminated through both glucuronidation and sulfation; however, the enzymes necessary for the former are essentially absent in fetal life. Therefore, the researchers sought to characterize isoforms of sulfation enzymes (for example, SULT1A1, SULT1A2 and so on) in terms of expression in fetal and pediatric liver samples, along with variability of successful acetaminophen biotransformation. They found that SULT1A1, SULT1A3/4, SULT1E1 and SULT2A1 were all detected in fetal liver, consistent with previous findings^{16, 17, 18, 19} and that SULT1A3/4 stood out as the most active sulfation enzyme during fetal life. They conclude that this would be a prime candidate gene for pharmacogenetic analyses of acetaminophen-associated gastroschisis. It is noteworthy that this study compared fetal and pediatric liver enzymes and found numerous differences between pre- and postnatal enzyme profiles.¹⁵

Little data have been reported regarding the presence and rates of different isoforms of these enzymes in different populations, let alone how these variations may be involved in actual drug response and toxicity. In 2005, recognizing the importance of CYP3A7 in fetal liver, Rodriguez-Antona et al.²⁰ reported a CYP3A7 variant that leads to increased catalytic activity (CYP3A7^*2). Interestingly, the frequency of CYP3A7^*2 was 8, 17, 28 and 62% in white, Saudi Arabian, Chinese and Tanzanian individuals, respectively. It is still unclear whether this mutation has a role in any human birth defects, but future studies evaluating the prevalence of specific genotypes in a variety of patient groups may one day clarify population differences in disease prevalence and drug efficacy.

Whereas fetal pharmacogenomics concerns itself primarily with in utero drug exposure, neonatal pharmacogenomics, in addition, addresses drug exposure through breastmilk. Similar to changes during fetal life, neonatal growth may also involve shifts in gene expression based on developmental stage; therefore, genes that may be involved in drug response at a given time in neonatal development may not be relevant in another. The following sections outline some areas in which neonatal pharmacogenomics have already made interesting progress.

Opioids

One of the first clinical depictions of the importance of neonatal pharmacogenomics followed the description of adverse effects observed in neonates who were breast-fed by mothers who received codeine. Codeine is normally converted to morphine through O-demethylation by CYP2D6. The conversion of codeine into norcodeine by CYP3A4 and into codeine-6-glucuronide by glucuronidation usually represents 80% of codeine clearance, whereas conversion of codeine into morphine by CYP2D6 represents only 10% of codeine clearance.²¹ Morphine is further metabolized into morphine-6-glucuronide and into morphine-3-glucuronide. Morphine and morphine-6-glucuronide have opioid activity. Glucuronides are eliminated by the kidney and are thus susceptible to accumulation in cases of acute renal failure.

CYP2D6 is one of the most studied CYP enzymes, because it metabolizes approximately 25% of all medications in the human liver.²² Approximately 7 to 10% of Caucasians lack any CYP2D6 activity because of deletions and frameshift or splice-site mutations of the gene. In contrast, approximately 1 to 3% of Middle Europeans, and up to 29% of Ethiopians show gene duplications, leading to elevated so-called ultra-rapid metabolism rates.²²

In 2004, Gasche et al.²¹ described an episode of codeine intoxication in a male patient who received codeine because of chronic pain related with chronic lymphocytic leukemia. A duplication of the CYP2D6 gene indicated that the patient was an ultra-metabolizer of codeine. In that case, the codeine toxicity was further compounded by the fact that the patient received clarithromycin and voriconazole antibiotics, which are known inhibitors of CYP3A4, thus reducing codeine clearance.²¹ In 2006, Koren et al.²³ reported a case of neonatal demise secondary to maternal codeine use after a vaginal delivery. Maternal genotyping of CYP2D6 revealed a CYP2D6^*2 × 2 duplication, classifying the mother as an ultra-rapid metabolizer. Elevated morphine concentrations were also found in both the serum of the mother and neonatal post-mortem blood.

An analog of codeine, tramadol hydrochloride, is a racemic mixture of two enantiomers. CYP2D6 and to a lesser extent CYP2B6 catalyze the O-demethylation of tramadol. In 2005, Allegaert et al.²⁴ suggested that CYP2D6 polymorphisms may explain the observed variability in O-demethylation activity of tramadol in the first months of life, and in 2008 the same group reported their data on the effects of CYP2D6 polymorphisms on tramadol O-demethylation in critically ill neonates and infants.²⁵ Recently, a Canadian case–control study by Madadi et al.²⁶ linked maternal CYP2D6 and UGT2B7 ultra-rapid metabolizer genotypes with marked post-partum fetal central nervous system depression and even demise. Moreover, mothers of symptomatic infants were 8 times more likely to have the ultra-rapid metabolizer genotype combinations than the average expected in the study's Western European population.

Selective serotonin reuptake inhibitors

A recent analysis found the prevalence of major depression to be as high as 5%, and that of major and minor depression to be as high as 10%.²⁷ Selective serotonin reuptake inhibitors (SSRIs) are commonly used antidepressants that act by inhibiting serotonin reuptake in the synaptic cleft. Medications in this group include fluoxetine, paroxetine, sertraline, fluvoxamine, citalopram and escitalopram. At higher doses, paroxetine and sertraline also block dopamine reuptake, which may contribute to their antidepressant action.²⁸ Venlafaxine is a combined serotonin–norepinephrine reuptake inhibitor. Common side effects of these medications include insomnia and agitation, nausea and other gastrointestinal effects, sexual dysfunction and weight gain.²⁸ Recently, multiple studies have correlated SSRI use during pregnancy with adverse neonatal effects, including neonatal respiratory distress, persistent pulmonary hypertension, jaundice, feeding problems, abnormal movements and tonus abnormalities, birth weight below 10% and even congenital cardiac disease (paroxetine only).^{29, 30, 31}

In 2008, a study by Oberlander et al.³² analyzed whether neonatal effects of SSRIs are related to genotypes for the serotonin transporter (SLC6A4) promoter; specifically, whether effected neonates were carriers of the short (s) or long (l) SLC6A4 alleles. The study suggested that prenatal SSRI exposure was associated with adverse neonatal outcomes and these effects were moderated by infant SLC6A4 genotypes. Moreover, the relationships between polymorphisms and specific outcomes varied during the neonatal period, suggesting that beyond apparent gene-medication interactions, multiple mechanisms contribute to the adverse neonatal outcomes after prenatal SSRI exposure.³²

However, studies exploring the role of maternal CYP enzymes in SSRI metabolism have not found similar correlations. In their study of 25 breastfeeding mothers who were treated with citalopram, sertraline, paroxetine, fluoxetine and venlafaxine, Berle et al.³³ evaluated multiple variations in maternal CYP2D6 and CYP2C19 genotypes, including genotypes associated with ‘poor metabolizing’. They found that no specific genotypes correlated with variations in excreted medications in breast milk and all of the infant's serum levels were either undetectable or low.

Sepsis

A major area that is currently under investigation is that of genetic susceptibility to neonatal sepsis. In 2003, Carcillo et al.³⁴ described reduced CYP-mediated drug metabolism in children with sepsis-induced multiple organ failure. As cytokines, such as interleukin-6 and nitric oxide production, increase in children with sepsis, and as such cytokines have been shown to reduce CYP450 activity both in vivo and in vitro, it was hypothesized that the degree of CYP450 inhibition is related in part with the degree of inflammation and organ failure. Children aged 1 day to 18 years, with and without sepsis, were enrolled and administered antipyrine through nasogastric tube, and antipyrine elimination rate, elimination half-life and apparent oral clearance were measured and calculated. As antipyrine metabolism was lower in children with sepsis, especially those with multiple organ failure, the researchers concluded that CYP-mediated drug metabolizing enzyme activity is markedly decreased in such circumstances.

Although a promising area, no studies have further correlated specific CYP enzymes with either sepsis susceptibility or antibiotic response. In their review of this topic, Del Vecchio et al.³⁵ suggest that in light of the growth continuum of neonates on a molecular, cellular and tissue level, further research in drug transport, drug biotransformation, receptors and signal transduction processes is needed before the promise of pharmacogenetics and pharmacogenomics for rational therapeutics can be realized in such cases.

Antineoplastic agents

Various drug-metabolizing enzymes are responsible for the metabolic activation and inactivation of potential carcinogenic agents and drugs used in the treatment of childhood malignancies.³⁶ This complex interplay of specific enzymes and agents involves multiple phases. Phase I metabolism converts many compounds to reactive, electrophilic, water-soluble intermediates, some of which can damage DNA.³⁷ This metabolic step is carried by CYP enzymes, and allele variants in phase I genes, such as CYP1A1, have been shown to be associated with higher enzymatic induction and have been implicated in susceptibility to childhood acute lymphoblastic leukemia.³⁸ Phase II drug-metabolizing enzymes, such as N-acetyltransferases and glutathione S-transferases, are important modifiers that selectively target specific chemotherapeutic drugs and potential carcinogenic compounds. These enzymes also detoxify various reactive species by catalyzing the conjugation of potentially mutagenic electrophilic substrates to glutathione. A recent study³⁹ of 209 children with neuroblastoma, 64% of whom were under the age of 2 years, details specific N-acetyltransferase1 and glutathione S-transferase M1 genotypes that result in a more favorable outcome in patients treated with standard chemotherapeutic agents that target this malignancy.

Alcohol and nicotine

Prenatal alcohol exposure can result in significant fetal, neonatal and long-term effects, including growth restriction, mental retardation, facial dysmorphogenesis, cardiac anomalies and other structural malformations.⁴⁰ Previous studies have shown that fetal alcohol concentrations closely resemble that of the mother, whereas the activity of alcohol dehydrogenase in the fetal liver is <10% of that observed in the adult.⁴⁰ The fetus therefore relies mostly on the maternal hepatic transformation of alcohol. Several adult studies have shown that genetic variations in enzymes involved in alcohol metabolism, specifically in the alcohol dehydrogenase genes ADH1B and ALDH2, can alter both the effects and addictive properties of alcohol consumption.⁴¹ ADH1B encodes an enzyme that converts alcohol into acetaldehyde whereas ALDH2 converts acetaldehyde into acetate. The prevalence rates of ADH1B and ALDH2 genotypes and allele frequencies vary among ethnic groups and likely contribute to the variations in alcohol dependence. Another genetic variation in CYP2E1, the enzyme that metabolizes alcohol to acetaldehyde and acetaldehyde to acetic acid, has been associated with greater ethanol elimination rates, heavier drinking and increased alcoholism.⁴¹

Similar to alcohol exposure, cigarette smoking has been linked with serious prenatal, perinatal and long-term adverse effects, including low birth weight, premature membrane rupture, preterm birth, perinatal death and sudden infant death.⁴⁰ Approximately 80% of nicotine is metabolized to cotinine by CYP2A6.⁴¹ CYP2A6-mediated nicotine metabolism varies widely both between and within ethnic groups, and different alleles can lead to poor, slow, intermediate and increased enzymatic activity. Several studies have linked genotypic variations in CYP2A6 with differences in smoking behaviors.⁴¹ Unfortunately, no studies to date have addressed the potential roles of genetic variations in enzymes that metabolize alcohol and tobacco with adverse fetal and neonatal outcomes.

Table 3 summarizes the previously described maternal, fetal–placental and neonatal pharmacogenomic enzymes and their effects.

Currently assisting clinicians and researchers understand, develop and apply pharmacogenomics is the Pharmacogenomics Knowledge Base (PharmGKB) (www.pharmgkb.org), a non-profit scientific consortium supported by the National Institutes of Health that curates information about the relationships among drugs, diseases and genes, including their variations and gene products. Its mission is to catalyze research by developing, implementing and disseminating a public genotype–phenotype resource that is focused on pharmacogenomics and pharmacogenetics. This resource serves a broad community including geneticists, molecular biologists, pharmacologists, physicians, policy makers and the lay public. The PharmGKB is currently actively curating gene variants that may have key roles in developing human embryo, including those actively transcribed by the placenta and fetal liver, such as CYP3A7 and CYP2J2.

Drug classification

Fetal and neonatal pharmacogenomics has the potential to change the way in which the medical establishment is able to counsel antenatal drug exposure by recognizing that the risks of drug exposure may be uniquely stratified. Currently, we use a drug classification system devised by the Food and Drug Administration that indicates five categories ranging from A to X (see Table 4). The result is that patients and their caretakers often must make decisions based on limited human data and crude animal studies, whereas the major ‘middle’ category (C) represents little to no applicable evidence.

A classic example involves antiepileptic medications that are often important enough for the health of the mother that their continued use is weighed seriously during pregnancy, but at the same time, they are associated with major malformations, including congenital heart defects, neural tube defects, oral-facial clefts and other minor anomalies. However, only a relatively small percentage (4 to 12%) of patients exposed to anticonvulsants will manifest these symptoms. Pharmacogenomics may one day identify a subgroup of patients at high risk for developing these undesired effects. For example, after following families with phenytoin (antiepileptic) exposure, Van Dyke et al.⁴² describe certain families with multiple exposed but normal offspring, whereas other families had multiple affected children. This kind of information would greatly direct medical treatment and risk counseling in pregnancy populations.

Antenatal drug efficacy

Other than pre-existing maternal medical conditions necessitating pharmacological intervention, obstetricians (and other practitioners) use a variety of medications during the course of pregnancy to treat pregnancy-related conditions. Examples of this include tocolytics for preterm labor, anti-hypertensives for pre-eclampsia and antibiotics for chorioamnionitis. Unlike known teratogens, most of these medications are relatively safe, but pharmacogenomics may assist clinicians in identifying subsets of populations that may benefit from one drug rather than another. Although still speculative, knowing that a patient's contractions will improve with nifedipine rather than indocin, or that labetalol is preferred over hydralazine in the case of hypertension may one day result in individualized care and improved drug efficacy. This phenomenon has recently been highlighted in the non-pregnant adult population by The International Warfarin Pharmacogenetics Consortium.⁴³ In their study of 1009 patients, the group showed that a pharmacogenetic algorithm dramatically improves the recommended initial warfarin dose compared with either a clinical or fixed-dose approach.

Recognizing the limited data, research and resources allocated toward the study of pharmacology in pregnant women and their infants, the National Institute of Child Health and Human Development recently developed the Obstetric-Fetal Pharmacology Research Unit intended to provide the ‘expert infrastructure needed to test therapeutic drugs during pregnancy’. Several studies are already underway, including ones evaluating glyburide for gestational diabetes and 17-α-hydroxyprogesterone caproate for preterm birth. Undoubtedly, pharmacogenomics will aid this research consortium and others to achieve their goals of conducting ‘a whole new generation of safe, technically sophisticated, and complex studies that will help clinicians protect the health of women, while improving birth outcomes and reducing infant mortality’.

Non-invasive maternal and fetal pharmacogenomic testing

Since the discovery of fetal cells in maternal blood several decades ago, investigators have used both intact fetal cells and cell-free fetal DNA in the maternal circulation as potential targets for non-invasive prenatal diagnosis.⁴⁴ Although much of the attention has focused on anueploidy detection, the development of advanced molecular tools, including digital PCR, microarray comparative genomic hybridization and mass sequencers, have opened the door to the possibility of complete non-invasive fetal genotyping while still in utero. As <1% of all neonates born to women under the age of 35 years are at risk for chromosomal abnormalities, fetal pharmacogenomic profiling may one day be as important a target as aneuploidy, if not even more so, for non-invasive prenatal diagnosis.

As the population ages, the prevalence of chronic diseases increases, and the average age of pregnancy goes up, practitioners can expect to see a greater number of patients with medical conditions necessitating chronic drug use: hypertension, diabetes, thyroid disorders and chronic infections are only some of many examples. Daily or even temporary medication use by expectant mothers is an area of scrutiny (and anxiety) because of the fear of potential adverse fetal and neonatal effects. As a result, many expectant mothers and their medical caretakers need to weigh the balance between risk and benefit of antepartum drug exposure.¹²

Although research in the area of drug metabolism in pregnancy is rapidly evolving, physicians are often hindered by limited knowledge of pathophysiological mechanisms that link a specific drug and adverse outcomes, as well as by a current drug classification system that rates a drug's safety in pregnancy based on limited animal data and weak associative studies. In today's world of rapid genetic testing, this approach seems increasingly archaic. Although still a developing field, research in maternal–fetal–neonatal pharmacogenomics will undoubtedly not only improve our overall understanding of fetal and neonatal physiology and development, but also help us to stratify patients based on individual drug risk, susceptibility and response, ultimately guiding better care for all women, families and children.

We thank Dr Brian Mercer for his insightful comments and suggestions. We also thank Dorit Berlin and her colleagues at the pharmGKB for their assistance.

Affiliations

1. Department of Obstetrics and Gynecology, Stanford University School of Medicine, Stanford, CA, USA

   • Y J Blumenfeld
   •  & Y Y El-Sayed

2. Yale University School of Medicine, New Haven, CT, USA

   • M F Reynolds-May

3. Department of Bioengineering, Stanford University, Department of Genetics, Stanford University School of Medicine, Stanford, CA, USA

   • R B Altman

Competing interests

The authors declare no conflict of interest.

Corresponding author

Correspondence to Y J Blumenfeld.

To obtain permission to re-use content from this article visit RightsLink.

Publication history

DOI

https://doi.org/10.1038/jp.2009.183