Abstract

The 21st century has been billed as the era of “precision/personalized medicine.” Genetic
investigation of clinical syndromes may guide therapy as well as reveal previously unknown biological
or pharmacological pathways that may result in novel drug therapies. Several clinical issues in
obstetrics and obstetric anesthesiology have been targets for genetic investigations. These include
evaluation of the genetic effects on preterm labor and the progression of labor, spinal anesthesia-
induced hypotension and the response to medications used to treat hypotension, and the effect of
gene variants on pain and analgesic responses. Most studies have examined specific single nucleotide
polymorphisms. Findings have revealed modest effects of genetic variation without tangible impact
on current clinical practice. Over the next decade, increased availability of whole exome and genome
sequencing, epigenetics, large genetic databases, computational biology and other information
technology, and more rapid methods of real-time genotyping may increase the impact of genetics in
the clinical arena of obstetrics and obstetric anesthesia.

Introduction

The National Institutes of Health’s Pharmacogenetics Research Network began a Pharmacogenomics

Knowledge Base (PharmGKB) in 2000 with the goal “to collect, encode, and disseminate knowledge
about the impact of human genetic variations on drug response, curate primary genotype and
phenotype data, annotate gene variants and gene-drug-disease relationships via literature review,
and summarize important pharmacogenetic genes and drug pathways” [1]. In line with these goals,
the Clinical Pharmacogenetics Implementation Consortium (CPIC)was launched in 2011 to establish
clinical recommendations for drug dosing based on genetic testing, providing information on
genotype-test interpretation and on the scientific evidence supporting the validity of these tests [2].

This review article aims to summarize the most recent evidence regarding genetic contributions to
clinical issues encountered by, and of interest to, obstetric anesthesiologists (Table 1). These include
genetic effects on the following conditions: (1) preterm labor and labor progression and delivery, (2)
hemodynamic and vasopressor response during spinal anesthesia for cesarean delivery, (3) labor pain
and analgesia, (4) post-cesarean pain and opioid response, and (5) fetal and neonatal effects of
maternal medication. We have selected these clinical phenotypes because of their relevance to the
practice of obstetric anesthesiologists, obstetricians and perinatologists. Numerous genes
undoubtedly influence the overall experience of labor and delivery; we focus here on the most well-
studied genes in the clinical context of obstetric anesthesiology and perinatology. While the direct
clinical applications of pharmacogenetic association studies in our field may still seem remote, lessons
learned from studies conducted in the last two decades have been enlightening.

<A>Preterm labor, labor progression and delivery

The contribution of genetic variants to preterm labor, and the progression and outcome of term labor,
including mode of delivery, have been the focus of several clinical studies. The phenotype of “preterm
birth” is defined as having 5 components: (1) maternal conditions present before delivery, (2) fetal
conditions present before delivery, (3) placental pathologic conditions, (4) signs of the initiation of
parturition, and (5) the pathway to delivery [3]. Genetic variants could in theory exert effects on any
of these five components, therefore the nature of and contribution of any genetic variant to overall
risk or outcome is difficult to quantify.

The b2-adrenergic receptor gene (ADRB2), in particular two common polymorphisms found in strong
linkage disequilibrium, p.16Arg/Gly and p.27Gln/Glu, have been extensively evaluated in a number of
anesthesia-related studies with relevant clinical effects described in the context of airway
responsiveness [4] and cardiac perioperative outcomes [5]. As b- adrenoceptors exist on the human
uterus, mediating uterine smooth muscle relaxation, and because they were a major pharmacologic
target for tocolysis in preterm labor or other forms of uterine hypercontractility, this genetic variant
was an obvious early candidate gene for studies of genetic effects on preterm labor and delivery.

In a small association study in a North-American cohort, homozygosity for Arg16 of ADRB2 was shown
to confer protection from preterm delivery [6]. A subsequent case-control study in a Hungarian
population also demonstrated a strong protective effect of Arg16 homozygosity, as no woman with
spontaneous preterm birth or rupture of membranes was Arg16 homozygous in this cohort of women,
20% of whom displayed this genotype [7]. A Turkish study suggested that the Glu27 variant of ADRB2
increases the risk for preterm delivery [8]. Thereafter, in a cohort of Swiss women diagnosed with
idiopathic preterm labor between 24 and 34 weeks’ gestation, hexoprenaline, a b2-adrenergic agonist
administered for tocolysis, was found to be more effective among women homozygous for Arg16
homozygous, demonstrating a true pharmacogenetic effect [9].

In a large prospective study evaluating the impact of ADRB2 genotype/haplotype on preterm labor, a
sub-analysis of North American women with term or near-term labor (> 34 weeks) confirmed the
hypothesis that Arg16 homozygosity of ADRB2 is associated with a slower rate of labor progress
compared with women with all other genotypes [10]; women Arg16Arg – Gln27Gln (double
homozygous) progressed at the slowest rate of active labor (‘slow haplotype’), and women Gly16Gly
– Glu27Glu (double homozygous) progressed with the fastest cervical rate (‘fast haplotype’).

Another study from the same North American group used a previously validated nonlinear mixed
effects model (NONMEM) of labor progress to predict labor progress and pain during the first stage of
labor in nulliparous women, according to ADRB2 and OPRM1 (m-opioid receptor gene). Women with
ADRB2 Gln27 homozygosity had a slower transition to active phase labor, resulting in prolonged labor,
compared to women with the other genotypes [11]. Genotype was strongly correlated with ethnicity
(e.g., no Asian women included in the study was Glu27 homozygous), and labor was indeed slower
among Asians. The study was underpowered to detect an effect of OPRM1 (rs1799971, codon 40) on
labor pain, but identified cold sensitivity as a pain modality that may predict the intensity of labor
pain.

Using the same model to evaluate the progression of labor, ADRB2, COMT (catechol- O-
methyltransferase gene) and OXTR (oxytocin receptor gene) were examined in a Saudi Arabian cohort
[12]. One variant of COMT (rs4633), along with one of OXTR (rs53576), were associated with increased
duration of the latent phase of the first stage of labor in nulliparous women, resulting in longer labors
(prolonged approximately 5 hours for COMT and 2 hours for OXTR). These findings among Saudi
women contrast with results from other studies that found no effect of OXTR single nucleotide
polymorphisms (SNPs) on the duration of the first stage of labor [11] or risk of dystocia [13]. In
addition, ADRB2 genotype/haplotype was not

associated with labor progress once demographic factors were taken into account [12]. Data
combining all the genetic variants (joined allelic combination) for each woman in the cohort were not
reported, therefore assessing the contribution of specific haplotypes on labor progress was not
possible.

Taken together, studies evaluating obstetrical outcomes based on ADRB2 genotype suggest that
women carrying Arg16 and/or Gln27 are conferred protection from preterm delivery and have a more
‘quiescent’ uterus, resulting in a slower progression and prolonged labor duration. In addition to
ethnicity and maternal weight, factors influencing labor progress are multiple, and ADRB2 is only one
of many possible candidate genes that may influence labor progression. Therefore, at the current
time, evaluating women’s genetic profiles is unlikely to contribute in a significant manner to clinical
decision-making or risk assessment.

At the opposite extreme of the clinical spectrum from preterm labor (the “overexcitable uterus”), a
genome-wide scan study in Sweden has identified at least 6 loci that appear to contribute to the risk
of women requiring a cesarean delivery for dystocia (the “poorly contractile uterus”) [13]; further
prospective studies will be needed to validate or refute these findings.

<A>Hemodynamics and vasopressor response during spinal anesthesia for cesarean delivery

The response to vasopressor agents (α- and β-adrenergic agonists) to prevent and manage maternal
hypotension following neuraxial anesthesia for elective cesarean delivery has been extensively studied
[14]. To date, five studies have evaluated the effect of ADRB2 genotype/haplotype on maternal
hemodynamic and vasopressor requirement during cesarean delivery under spinal anesthesia [15-19].
In the first of these studies in a North-American cohort, the incidence and severity of maternal
hypotension after spinal anesthesia for cesarean delivery, and the response to treatment, was
affected by ADRB2 genotype/haplotype [15]. Women with either Gly16Gly - Gln27Glu or Gly16Gly –
Glu27Glu (double homozygous) haplotype were found to require significantly less vasopressor
(ephedrine) for treatment of hypotension during spinal anesthesia. These two “protective” haplotypes
associated with less hypotension are relatively common in Caucasians (20%). In contrast, in a smaller
Brazilian cohort, Arg16 homozygosity appeared to confer protection from hypotension, and ephedrine
requirements were lower in women with that genotype [16]. In a follow-up study by the same authors
that included data from the same cohort, the double homozygous Arg16Arg - Gln27Gln haplotype
appeared again to reduce the incidence of hypotension and need for ephedrine, although most of the
difference in hypotension incidence occurred after delivery (15-30 minutes after spinal injection) [18].
The influence of spinal anesthesia itself, versus blood loss, oxytocin or other post-delivery factors, was
unclear.

In a cohort of Asian women [17], the 2 haplotypes previously found to be associated with lower
ephedrine requirements in the North-American cohort [15]were under-represented and ADRB2
genotype/haplotype was not found to significantly influence maternal ephedrine requirements [17].

A follow-up study from the North American group, using a more contemporary vasopressor regimen
with a phenylephrine infusion, again evaluated spinal hypotension and phenylephrine dosing during
cesarean delivery [19]. As the effect of phenylephrine, an α1- adrenergic agonist, is unlikely to be
directly affected by ADRB2 genotype, any differences in phenylephrine requirements based on ADRB2
are likely an effect on blood pressure regulation in response to spinal anesthesia, rather than a specific
pharmacogenetic effect on spinal anesthesiainduced hypotension. In the adjusted analysis, Arg16
homozygosity was associated with moderately greater phenylephrine doses than other genotypes.

Taken together, the two studies from the North American group suggest that the pharmacologic effect
of ephedrine in the setting of treatment for spinal anesthesia-induced hypotension is altered by
ADRB2 genotype because the dose of phenylephrine needed to maintain blood pressure after spinal
anesthesia [19] is less affected than the ephedrine dose by ADRB2 genotype [15]. In both North
American studies, in contrast to the Brazilian cohort, Arg16 homozygous patients received more drug
(ephedrine or phenylephrine), suggesting that Arg16 homozygosity predisposes to hypotension with
spinal anesthesia, and therefore, higher vasopressor requirements. Historically, ephedrine was the
drug of choice to treat spinal anesthesia-induced hypotension during pregnancy; although the current
clinical practice is to prevent hypotension with phenylephrine, administered either as a bolus or a
continuous infusion. Nonetheless, pharmacogenetic effects may explain why two decades of multiple
studies trying to define a single optimal strategy (fluid loading, ephedrine or phenylephrine) to prevent
or treat hypotension during spinal anesthesia for cesarean delivery have failed to identify that one
regimen that ‘fits all.’

<A>Labor pain and analgesia

There is an extensive body of literature on the genetics of pain and response to opioids [20], however,
findings are not particularly congruent across studies. Specifically, genetic associations seem to vary
when examining different pain phenotypes. Indeed, studies evaluating genetic influences on
experimental pain in healthy volunteers, oral opioid therapy in cancer patients, postoperative opioid
consumption or neuraxial labor analgesia effects have reported divergent, sometimes opposite,
effects. A likely explanation for these divergent results is that different pain modalities, such as the
visceral component of labor pain, are inherently distinct from the predominantly somatic component
of postsurgical pain, and that different opioids, delivered via different routes, may affect receptors
and pathways in different parts of the nervous system, and undergo metabolism by different
metabolic pathways with different pharmacogenetic influences.

The gene coding for the μ-opioid receptor, OPRM1, is one of the most widely studied genes in the
context of pain, and obstetric pain in particular [21,22]. Genetic variants of OPRM1 have been
evaluated in numerous clinical acute and chronic pain settings, as well as in patients with alcohol and
opioid addiction. An OPRM1 single nucleotide polymorphism, in which adenine is substituted by
guanine at nucleotide position 118 (resulting in the replacement of asparagine by aspartate at codon
40 of OPMR), is the most well studied SNP in the setting of pain and opioid analgesia. The frequency
of the G118 allele varies according to race/ethnicity, occurring in approximately 30% among non-
Hispanic Caucasians [23], 60% to 80% among Asians [24], and likely lower among Hispanics and African
Americans.

The first clinical trial designed to study the effect of genetic influences on labor pain and analgesic
response to neuraxial opioids evaluated the A118G genetic variant of OPRM1 in a Swiss cohort of
healthy nulliparous women [25]. Women were enrolled in the third trimester and based on
genotyping, were classified as homozygous A118, or heterozygous/homozygous G118. If enrolled
women requested early labor neuraxial analgesia (< 6 cm cervical dilation), their response to
intrathecal fentanyl, administered as a component of combined-spinal epidural (CSE) analgesia, was
assessed. The study included two separate trials to estimate the median effective dose (ED50) of
intrathecal fentanyl based on genotype, defined as the dose of fentanyl providing complete analgesia
(verbal rating scale score ≤ 1/10, 0-10 point scale) for 60 minutes in 50% of women. Using a sequential
allocation design, the ED50 for fentanyl was greater (by a factor of 1.5) among A118 homozygotes
compared to women with the two other genotypes. In a separate trial using a random-dose allocation
method, this finding was confirmed with very similar results; the ED50 was 2-fold greater among A118
homozygotes. In addition, cervical dilatation at the time of labor analgesia request was significantly
lower among A118 homozygotes, indicating that these women requested analgesia earlier in labor
than women carrying at least one G118 allele. Taken together, these findings suggest that A118
homozygotes require a significantly higher dose of intrathecal fentanyl for early labor analgesia. This
is an important finding for clinicians, because the spinal dose cannot be titrated; a dose must be
chosen before the effects are assessed, and the adverse effect profile of intrathecal fentanyl is dose-
dependent (e.g., nausea, pruritus). Therefore, reducing the dose in women with lower dose-
requirements has the potential to improve clinical outcomes. The higher dose requirement would
apply to approximately 30% of Caucasian women and even higher percentage of Asian women.
Following this first study, the investigators sought to determine whether A118G variant of OPRM1 also
influences the duration of intrathecal fentanyl analgesia for early labor analgesia [26]. This prospective
observational trial was performed in a cohort of North American women. As in the previous study,
women were classified in one of two groups based on OPRM1 genotyping. The duration of intrathecal
fentanyl analgesia was defined as the time elapsed between administration of the intrathecal fentanyl
dose and the next parturient request for analgesia (administered via the epidural catheter). There
were no significant differences between genetic groups in the duration of intrathecal fentanyl
analgesia in early labor. A possible explanation for these negative results is that all women received a
high dose of fentanyl (25 μg), and therefore, a subtle genetic influence may have been masked by this
large dose.

As a follow-up to the study evaluating the effect of A118G variant of OPRM1 on the response to
intrathecal fentanyl for early labor analgesia, the response to epidural sufentanil was evaluated in a
cohort of healthy nulliparous Italian women using an up-down sequential allocation study design [27].
Genotyping performed between 35-37 weeks’ gestation and women classified as A118 homozygotes
or heterozygous/homozygous G118. Similar to the findings with intrathecal fentanyl, the ED50 was
significantly greater (by 25%) among women A118 homozygotes compared with women with the two
other genotypes. Based on findings from these three studies in healthy nulliparous women delivering
with early neuraxial analgesia, it can be reasonably concluded that the A118G variant of OPRM1
influences the potency of neuraxial lipophilic opioids (fentanyl or sufentanil) for labor analgesia, with
lower dose-requirements in women carrying the G118 allele, but this SNP does not influence the
duration of analgesia.

A group of Swedish investigators conducted an observational study to investigate whether the A118G
variant of OPRM1 influences the cervical dilation at the time of arrival in the labor and delivery unit,
as a surrogate to estimate the progression of painful labor, and the requests for different analgesic
modalities during labor [28]. Women were categorized as A118 homozygotes (81%), A118G
heterozygotes (18%), and G118 homozygotes (2%). There was no association of cervical dilation on
arrival at the delivery unit or use of any type of analgesia (epidural or second-line analgesia) during
labor with A118G genotype. These authors concluded that this genotype is unlikely to be of major
importance in predicting pain related behaviors during labor.

The guanosine triphosphate cyclohydrolase gene (GCH1) displays up to 15 SNPs and has been
associated with reduced pain sensitivity in a variety of pain settings. Pain behaviors during labor and
delivery were assessed in the same cohort of Swedish women [28], with genotyping of “pain-
protective” SNPs of the GCH1 gene [29]. Women were classified as homozygous (2%), heterozygous
(27%) and non-carriers (71%) of pain-related SNPs. Homozygous carriers of SNPs associated with ‘pain-
protection’ were admitted to the labor and delivery unit in more advanced labor (measured by cervical
dilation) than heterozygous carriers and non-carriers. The authors concluded that GCH1 may influence
the pain experience during the initial stage of labor. Unfortunately, because the epidural analgesia
rate was relatively low (approximately 20%) and epidural analgesia services were not available 24
hours per day in the institution where this study was performed, any effect of GHC1 variants, or
OPRM1, on the need for or request for analgesia was difficult to assess.

Interestingly, findings in laboring women disagree with other studies examining the influence of
A118G genotype of OPRM1 and opioid analgesia on pain [30]. These studies include studies of
postoperative intravenous fentanyl consumption [31-33], spinal morphine for postcesarean pain
[24,26], intravenous morphine for postoperative pain [34,35], or oral morphine for chronic cancer pain
[36-38]. Potential explanations for such discrepant results are that labor pain is different from that
experienced in other clinical settings (e.g., experimental pain, postoperative pain, chronic pain), or
perhaps the response to systemic compared with neuraxial opioids is affected differently by OPRM1
genotype because of differences in receptor mechanism and coupling in different areas of the
neurological system, or secondary/linked genetic changes related to neural connections and pain
transmission.

The possible effect of polymorphisms of COMT and OPRM1 on analgesia from intravenous fentanyl
for labor analgesia was examined in one study [39]. The study did not show any specific effects of
either gene on the success of fentanyl analgesia and was underpowered to find effects of
combinations of genotypes. While the major finding of this study was that intravenous fentanyl is not
effective for labor analgesia, the findings also suggest that one should not expect major differences in
intravenous fentanyl analgesia efficacy based on genetics; no particular genetic population seems to
benefit significantly more than others from such therapy. Genetic influences, if any, are probably
modest in this context.

<A>Postcesarean pain and response to post-operative analgesia

Three studies have evaluated the influence of A118G genotype of OPRM1 on postcesarean analgesia
after spinal or intravenous morphine [24,26,40]. In a North American cohort, need for postcesarean
rescue analgesia after a spinal anesthesia bupivacaine and morphine 150 μg and fentanyl 15 μg was
examined by individual patient genotype; there was no difference in the duration of spinal morphine
analgesia or need for analgesic supplementation over 72 hours in women carrying the G118 allele [26].
The incidence of nausea was similar between groups; however, pruritus was less frequent in carriers
of the G118 variant allele during the first 24 hours. The lack of major differences between genotypic
groups exposed to a single intrathecal morphine dose does not rule out a significant difference in dose
response, as spinal morphine 150 μg is near the top of the dose-response curve.

In two studies among Asian women undergoing cesarean deliveries with spinal anesthesia with spinal
morphine 100 μg [24, 40], carriers of the G118 variant allele exhibited greater 24-hours postoperative
morphine consumption of morphine, administered with intravenous patient-controlled analgesia
(PCA), a result opposite to that reported in North American and European women receiving neuraxial
fentanyl and sufentanil for laboranalgesia [25, 27]. In Singaporean women of Chinese ethnicity, 24-
hour postoperative morphine intravenous PCA consumption was significantly lower in A118
homozygotes compared with the 2 other genotypes [24]. Distribution of morphine use over time
demonstrated that most morphine use occurred in the postanesthesia care unit (PACU) during the
first hours after spinal anesthesia. This early intravenous PCA consumption may have reflected lack of
analgesia upon arrival in the PACU (the peak effect of intrathecal morphine is several hours). The
overall incidence of nausea was low; nonetheless it was significantly higher in A118 homozygotes than
the other two groups. Because pain scores were lower in A118 homozygotes, nausea was unlikely to
have discouraged women from self-administering morphine using PCA.

In a second publication by the same authors, Asian women were evaluated (n=617 Chinese, n=241
Malays and n=136 Indians) after receiving intrathecal morphine 100 μg as a component of spinal
anesthesia [40]. The overall genotype distribution was 39% A118 homozygotes, 44% 118A/G
heterozygotes and 17% G118 homozygotes, with significant differences among ethnic groups (G118
homozygosity more prevalent among Malays and Indians). The authors reported large interindividual
variability; 65 women did not selfadminister any intravenous morphine, 129 administered only one
dose, while another 122 administered 2 doses. Total intravenous morphine use was significantly
higher in G118 homozygotes, and the incidence of nausea was again lower in women with this
genotype. In a multiple regression analysis, the most important variable contributing to morphine use
was the maximum pain score, followed by ethnicity and A118G genotype. After correction for OPRM1
genotype, ethnicity was still a significant contributing factor; Indian women reported higher pain
scores and used higher doses of intravenous morphine. It should be noted that in both studies the
actual amount of intravenous morphine used by women was low and quite variable, and the mean
differences (a few milligrams) among groups may not be clinically significant.

The apparent discrepancy between the North American study reporting no effect of A118G genotype
of OPRM1 on spinal morphine analgesia and findings in the Asian cohort may be explained by
differences in study design and primary outcomes. In the studies evaluating Asian women, the spinal
solution did not include fentanyl; it is possible that onset morphone analgesia occurred after women
arrived in the PACU. Because women were allowed intravenous PCA morphine as the initial rescue
analgesic (rather than ibuprofen as in North American study [26]), such a design was more likely to
detect an effect of A118G genotype on intravenous analgesia response rather than an effect of spinal
analgesia, at least for the early postoperative period doses. Another explanation may be that OPRM1
genotype may interact differently with opioid analgesia in different ethnic groups, as is suggested by
the recent finding of an ethnicity-dependent genetic association for the A118G genotype of OPRM1
described in volunteers evaluated with an experimental model of pain [41].

Lastly, a study on postcesarean epidural analgesia in an Asian population failed to show a difference
in medication requirements by OPRM1 genotype [42]. In this study, an epidural solution of ropivacaine
with sufentanil was used for postcesarean delivery analgesia. No differences in epidural dose
requirements were found between A118G genotypic groups; however, the most the medication was
administered via a background infusion rather than in response to pain or analgesic response, limiting
the ability of the study to identify differences between groups.

In a pilot study, the relationship between postcesarean pain scores, codeine consumption and side
effects, and CYP2D6 genotype were evaluated [43]. Codeine is a prodrug, and requires O-
demethylation catalyzed by CYP2D6 for conversion into morphine for analgesic effect; this metabolic
pathway accounts for 10% of codeine clearance. The conversion of codeine into norcodeine by CYP3A4
and into codeine-6-glucuronide by glucuronidation represents approximately 80% of codeine
clearance. Morphine is further metabolized into morphine-6-glucuronide (M6G) and morphine-3-
glucuronide (M3G); both morphine and M6G display opioid activity. Individuals with an “ultra-rapid
metabolizer” (UM) phenotype have increased risk for respiratory depression with codeine, particularly
if CYP3A4 activity is inhibited by concomitant use of antibiotics or in the context of kidney dysfunction
[44]. In the pilot study, two “poor metabolizers” (PMs), as predicted by genotype, did not achieve
analgesia, and two of three UMs, expected to generate higher morphine concentrations, discontinued
codeine use because of adverse effects. In a larger study from the same group of investigators
examining the effect of CYP2D6, OPRM1, COMT, ABCB1 (ATP-binding cassette, sub-family B, member
1 gene) and UGT2B7 (UDP glucuronosyltransferase 2 family, polypeptide B7) on codeine analgesia,
only OPRM1 and UGT2B7 appeared to predict codeine use, and only in Caucasian women, but not
Asian women [45]. Although not all studies have shown that clinical outcomes are affected by
genotype, codeine is falling out of favor as a routine analgesic because of concerns about differences
in metabolism of this pro-drug, and variable transfer of morphine to the neonate during breastfeeding
(vide infra).

Finally, the risk of chronic pain after cesarean delivery has been the subject of several recent
publications [46-50]. Persistent pain (defined as pain 8 to 12 weeks postpartum) or chronic pain (pain
at 6 or 12 months postpartum) postcesarean delivery has been associated with SNPs of the ABCB1 in
an Asian cohort of women evaluated in Singapore [51]. These findings require prospective validation
with pain outcomes that include neuropathic pain descriptors across multiple ethnicities.
<A>Fetal and neonatal effects of maternal medication

<B>Neonatal acidosis after maternal ephedrine administration

Maternal ephedrine administration immediately prior to cesarean delivery (usually used to treat spinal
anesthesia-induced hypotension) has been shown to have a dose-dependent effect on lower neonatal
pH, presumably due to a fetal “hypermetabolic” β-adrenoceptor effect. Neonatal pH and the risk of
acidemia is greater in neonates delivered of mothers who receive ephedrine compared with
phenylephrine; thus ephedrine is no longer the first-line drug of choice for prevention and treatment
of maternal hypotension before cesarean delivery [52]. The hypothesis that the ADRB2 genotype of
the fetus/neonate directly influences the degree of neonatal acidemia in response to maternal
ephedrine administration was examined in a secondary analysis of a randomized controlled trial in
ethnic Chinese women comparing the placental transfer of ephedrine and phenylephrine [17]; there
was no effect of maternal ADRB2 genotype on ephedrine requirements for maintaining maternal
blood pressure. However, fetal ADRB2 genotype had a significant effect on neonatal umbilical artery
pH. The effect of ephedrine dose on neonatal pH was completely absent in Arg16 homozygous fetuses,
but clearly present in fetuses with other ADRB2 genotypes [17]. These findings provide interesting
insight on fetal acidemia and metabolic responses in neonates born to mothers who have received b-
adrenergic agonists (ephedrine and/or other b-stimulants prescribed for tocolysis or bronchodilation)
prior to delivery.

<B>Neonatal abstinence syndrome after maternal opioids

In a prospective cohort study conducted in the United States, the association between genetic variants
of OPRM1, COMT and ABCB1 and the risk for neonatal abstinence syndrome (NAS) in neonates
exposed in utero to methadone or buprenorphine was evaluated in motherinfant pairs [53]. Neonates
carrying the variant G118 allele of OPRM1 had a shorter length of hospital stay and were less likely to
require medication compared to A118 homozygous neonates. Neonates carrying one or two alleles of
the Val158 variant of COMT were also protected from a longer hospital stay and the need for
medication. The study was underpowered to evaluate the joined allelic combination of OPRM1 and
COMT on neonatal outcomes. There was no association of any ABCB1 polymorphism with neonatal
outcomes. This was the first study to evaluate the influence of fetal/neonatal genotype on neonatal
outcomes after in utero exposure to opioids. Another study from the same group of investigators
suggested an association of other opioid receptor gene variants (ĸ-opioid receptor (OPRK1), delta-
opioid receptor (OPRD1) and the prepronociceptin gene (PNOC), as well as OPRM1) on severity
(increased need for drug treatment) of NAS. This group also reported an epigenetic effect related to
NAS after maternal opioid exposure; severely affected neonates had increased levels of methylation
in the promoter region of OPRM1, suggesting decreased gene expression [54,55].

<B>Neonatal death after maternal codeine

Historically, codeine was a widely prescribed opioid analgesic because of the belief that this weak
opioid would not result in adverse outcomes. It has been considered a safe alternative to other opioids
for outpatient pain management, and is still available in some countries as an over-the-counter
medication, either alone or in combination with paracetamol (acetaminophen). Use of codeine has
included pediatric patients, although the United States Food and Drug Administration (FDA) banned
the use of codeine after tonsillectomy in 2013 [56], and large-scale evidence of efficacy is sparse and
has been challenged by a recent metaanalysis [57]. The most striking report of fatality after codeine
prescription was the the death of a breastfed 13-day-old neonate from what appears to be an opioid
overdose because his mother was taking codeine after childbirth [58]. Toxic blood levels of morphine
or its active metabolite morphine-6-glucuronide (M6G) may arise in mothers and neonates that are
CYP2D6 UM or extensive metabolizers (EM). The mother was categorized as a CYP2D6 UM and her
breast milk had a morphine concentration of 87 ng/mL; the typical range is 1.9 to 20.5 ng/mL at doses
of 60 mg codeine every 6 hours. The infant was categorized as a CYP2D6 EM and postmortem
toxicology tests using gas-chromatography mass spectrometry revealed an infant blood
concentrations of morphine of 70 ng/mL. A daily maternal codeine dose of 60 mg for postpartum pain
normally results in maximum morphine plasma concentrations of 2.2 ng/mL in breastfed neonates
[59]; neonates prescribed morphine for analgesia display serum morphine concentrations ranging
from 10 to 12 ng/mL [60].

Codeine and morphine clearance in breastfeeding mothers and their relation to CYP2D6 genotypes
have been extensively evaluated and discussed [61-69]. Since 2007, the FDA requires manufacturers
of prescription codeine products to state the risks of prescribing codeine to breastfeeding mothers in
the “Precautions” section of the drug label [70]. It has been suggested that in the absence of a clear
phenotype and no genotypic data for CYP2D6, codeine should be avoided in breastfeeding mothers.
An FDA-approved genetic test (AmpliChip CYP450: Roche Diagnostics, Palo Alto,CA, USA) is
commercially available to test genetic variants of CYP2D6 [71]. Overall, the level of evidence linking
gene variation (CYP2D6) to phenotype (increased biotransformation of codeine into morphine) is
significant; however, there is no randomized clinical trial assessing the benefits of genetic testing prior
to codeine therapy at large. In addition, while insufficient morphine formation following codeine
administration, resulting in failure of analgesia, can be predicted from genotyping, extremely high
morphine formation currently still requires combining genotyping with phenotyping data [72]. The
Clinical Pharmacogenetics Implementation Consortium published guidelines for genotyping and
codeine administration in 2014 that are based on a focused review and interpretation of the literature
by experts in the field [73]. Because of a wide interethnic genotype distribution, not only for CYP2D6,
it is important to identify the cohorts in which pharmacogenetic effects and genetic association have
been reported and avoid generalizing findings from one study to different populations with different
ethnic make-ups.

Summary

Accurate phenotyping is key in all genetic association studies. Furthermore, designing clinical studies
to assess the genetic contribution to clinical outcomes is remarkably challenging, a difficulty often
underestimated when interpreting results [74-76]. In addition, particularly when multiple genes are
evaluated, large sample sizes and appropriate statistical analysis are required to avoid spurious and
misconstrued findings [77].

Particularly relevant to the practice of obstetric anesthesiology, the genetic contribution to labor pain
is extremely complex and true pharmacogenetic associations to explain differences in analgesic
response are not straightforward. On the positive side, pharmacogenetic studies have demonstrated
that genetics do play a significant role in a variety of clinical conditions, including response to drugs,
highlighting that clinical studies should no longer be designed with the sole goal of identifying “the
one dose for one drug that will work for all.” Unfortunately, firm recommendations on tailoring opioid
regimens to an individual patient’s genetic profile are unlikely to become available in the near future,
other than for codeine. Genome-wide association studies or exome sequencing to examine well
known and yet-undiscovered genetic variants explaining various outlier phenotypes will be of interest.