SECTION III: SPECIAL POPULATIONS

INTRODUCTION

Reproductive and perinatal principles in toxicology are often derived from basic science studies and are applied cautiously to clinical practice. This chapter reviews several principles of reproductive medicine that have implications for toxicology, including the physiology of pregnancy and placental xenobiotic transfer, the effects of xenobiotics on the developing fetus and the neonate, and the management of poisoning or overdose in pregnant women.

One of the most dramatic effects of exposure to a xenobiotic during pregnancy is the birth of a child with congenital malformations. Teratology, the study of birth defects, has principally been concerned with the study of physical malformations. A broader view of teratology includes “developmental” teratogens—xenobiotics that induce structural malformations, metabolic or physiologic dysfunction, or psychological or behavioral alterations or deficits in the offspring, either at or after birth.²⁹¹ Only 4% to 6% of birth defects are attributable to known pharmaceuticals or occupational and environmental exposures.^291,302

Some reproductive effects of xenobiotics occur before conception. Female germ cells are formed in utero; adverse effects from xenobiotic exposure theoretically occur from the time of a woman’s own intrauterine development to the end of her reproductive years. An example of a xenobiotic that had both teratogenic and reproductive effects is diethylstilbestrol (DES), which was most notable for causing vaginal or cervical adenocarcinoma (or both) in some women who were exposed to DES in utero and which also had adverse effects on fertility and pregnancy outcome in the same cohort of women exposed in utero.^239,263

Men generally receive less attention with respect to reproductive risks. Male gametes are formed after puberty; only from that time on are they susceptible to xenobiotic injury. An example of a xenobiotic affecting male reproduction is dibromochloropropane, which reduces spermatogenesis and, consequently, fertility. In general, much less is known about the paternal contribution to teratogenesis.^63,130,152

Occupational exposures to xenobiotics are potentially important but are often poorly defined. In 2015, there were approximately 43 million women of reproductive age in the workforce.³³¹ Although approximately 90,000 chemicals are used commercially in the United States, only a few thousand of them have been specifically evaluated for reproductive toxicity. Many xenobiotics have teratogenic effects when tested in animal models, but there are relatively few well-defined human teratogens (Table 30–1).³⁰² Thus, most tested xenobiotics do not appear to present a human teratogenic risk, but most xenobiotics have not been tested. Some of the presumed safe xenobiotics have other reproductive, nonteratogenic toxicities. Several excellent reviews and online resources are available.^{39,106,266,302,325}

Another type of xenobiotic exposure for a pregnant woman is intentional overdose. Although a xenobiotic taken in overdose can have direct toxicity to the fetus, fetal toxicity frequently results from maternal pulmonary or hemodynamic compromise, further emphasizing the critical interdependency of the maternal–fetal dyad.

Xenobiotic exposures before and during pregnancy can have effects throughout gestation and also extend into and beyond the newborn period. In addition, the effects of xenobiotic administration in the perinatal period and the special case of delivering xenobiotics to an infant via breast milk deserve particular consideration.

PHYSIOLOGIC CHANGES DURING PREGNANCY THAT AFFECT DRUG PHARMACOKINETICS

Many physical and physiologic changes that occur during pregnancy affect both the absorption and distribution of xenobiotics in the pregnant woman and consequently potentially affect the amount of xenobiotics delivered to the fetus.²¹

Delayed gastric emptying, decreased gastrointestinal (GI) motility, and increased transit time through the GI tract occur during pregnancy. These changes result in delayed but more complete GI absorption of xenobiotics and, consequently, lower peak plasma concentrations which may be consequential in an adverse manner. Because blood flow to the skin and mucous membranes is increased, absorption from dermal exposure is increased. Similarly, absorption of inhaled xenobiotics is increased because of increased tidal volume and decreased residual lung volume.

An increased free xenobiotic concentration in the pregnant woman can be caused by several factors during the later stages of pregnancy a decreased plasma albumin, increased binding competition, and decreased hepatic biotransformation. Fat stores increase throughout pregnancy and are maximal at about 30 weeks; near term, free fatty acids are released, and along with them the lipophilic xenobiotics that have accumulated in the lipid compartment are released. The increased concentration of circulating free fatty acids leads to competition with circulating free xenobiotics for binding sites on albumin.

Other factors leading to decreased free xenobiotic concentrations, early in pregnancy, are increased fat stores, as well as the increased plasma and extracellular fluid volume, which lead to a greater volume of distribution. Increased renal blood flow and glomerular filtration may or can result in increased renal elimination.

Cardiac output increases throughout pregnancy, with the placenta receiving a gradually increasing proportion of total blood volume. Xenobiotic delivery to the placenta therefore increases over the course of pregnancy.

These processes interact dynamically, and it is difficult to predict their net effect. The concentrations of many xenobiotics, such as lithium, gentamicin, and carbamazepine, decrease during pregnancy even if the dose administered is unchanged.¹¹⁰

Although not specifically related to the physiologic changes occurring during pregnancy, the fetus can also be exposed to xenobiotics that had accumulated in adipose tissue before pregnancy. For example, typical retinoid malformations occurred in a baby born to a woman whose pregnancy began one year after she discontinued the use of etretinate (retinoic acid), a medication that is used to treat severe psoriasis, congenital ichthyosis acne, and several other disorders of keratinization which may have clinical implications.¹⁷⁹

XENOBIOTIC EXPOSURE IN PREGNANT WOMEN

Exposure to xenobiotics during pregnancy is common. At some time during pregnancy, as many as 90% of pregnant women take prescription or nonprescription medications other than vitamins, mineral supplements, and iron. The most commonly used are analgesics, antipyretics, antimicrobials, antihistamines, vitamin B₆ and antiemetics.^{73,74,198,250,310,328} In addition, use of caffeine, tobacco, and alcohol is common. Some pregnant women use xenobiotics to treat a preexisting chronic disease such as epilepsy or a newly diagnosed disease such as deep vein thrombosis. Many women use various prescription and nonprescription xenobiotics before recognizing that they are pregnant.

Following the international experience with thalidomide in the 1960s that identified a high frequency of congenital malformations associated with its use, the US Food and Drug Administration (FDA) decided there was a need for guidance to prescribers regarding the use of medications during pregnancy. In 1979, the FDA began to require pharmaceutical manufacturers to label their products with respect to use in pregnancy and introduced the ABCDX classification scheme familiar to most practitioners in the United States.³³² Classification systems similar to the former FDA system were developed in Sweden and Australia.^272,289

The intent of those classification systems was to inform practitioners about the nature of the available evidence regarding risk in pregnancy. However, the ABCDX system was criticized for a number of reasons—the manufacturer assigned the risk category, there was no requirement for additional research, and there was no requirement to update the label when new information became available.⁶

One of the most significant problems with the ABCDX system is the general impression among health care practitioners that the letter categories indicated teratogenic risk with a hierarchy of harmful effects and an equivalence of risk within each category.^82,296,297 For example, a category C medication was generally considered riskier than a category B medication in pregnancy even though category C was the default category for medications about which there was little or no specific human information available and for which the risk was unknown. Approximately 90% of medications were classified as category C.¹⁹³

In addition, there was significant discordance between the use-in-pregnancy labeling and the teratogenic risk as determined by clinical teratologists.^5,193 Manufacturers labeled certain medications as category X (contraindicated in pregnancy) even when there was only limited information associating the medication with any adverse fetal or neonatal effects. For example, oral contraceptives carried a category X classification even though they are not considered teratogenic; the category X was assigned because there was no indication for use of oral contraceptives in pregnancy.¹⁷⁶ Certain medications that carried a category D (high-risk) classification might have caused problems only at certain times during pregnancy or in particular circumstances. In several reports, approximately 6% of pregnant women were prescribed medications that were categorized as D or X and 1% were prescribed medications that are considered teratogenic.^5,13,193 Even medications that have been classified as category D or X sometimes had only a very low risk of teratogenicity or other adverse effects, and exposure to these xenobiotics, even during the first trimester, would not necessarily have been an indication to terminate a pregnancy.⁸⁶

The difficulty regarding appropriate drug labeling reflects many complex questions regarding the use of medications during pregnancy. How should animal data in general be evaluated? How should animal data be extrapolated to humans? How should the teratogenic risk be defined and quantified for any particular xenobiotic? How should the risk of not treating a particular disease be compared with the risk of using a particular medication to treat that disease? Finally, how should the answers to these questions be communicated to practitioners and the public?²⁹⁷

In an attempt to address many of these problems, the FDA organized a taskforce in 1998 to propose changes to the labeling of medications for use in pregnancy and lactation.^142,176 The changes, referred to as the “Pregnancy and Lactation Labeling Rule” (PLLR), were initially published in 2008, finalized in 2014, and took effect in 2015 with full implementation expected by 2020.³³⁴ The new PLLR requires “the removal of pregnancy categories A, B, C, D, and X from all human and prescription drug and product labeling” and specifies the content and format for new labels. This content requires (1) a concise description and estimate of the risk of structural teratogenesis, fetal or infant mortality, and impaired growth or physiologic function; (2) details of animal and human data, in particular information from registries, cohort studies, and case–control studies; and (3) clinical considerations such as risk of the disease (treated or untreated) versus risk of the medication, need for dose adjustment in pregnancy, adverse drug reactions specific to pregnancy, monitoring of drug use and effect, effects on labor and delivery, and neonatal complications. The PLLR also requires detailed information with regard to medication use during lactation (previously referred to as the “nursing mothers” section) as well as a new section that describes medication effects on both female and male fertility. There is also a requirement to update the label as new information becomes available that could make the label “inaccurate, false, or misleading.”

The new labels will be completely phased in by 2020. Drugs approved after 2015 must use the new labeling format. Drugs released after 2001 will have the newly formatted labels phased in over several years. Drugs approved and released prior to 2001 will not be required to use the new format but will be required to eliminate any reference to a previously assigned letter from the ABCDX system. Considering that the new FDA labels are being phased in over several years, the editors have decided to retain the ABCDX designation in other chapters for this edition.

Specific current information on individual xenobiotics can be obtained from local and regional teratogen information services and published books,^39,302 some of which also have online versions.^228,266,325 Motherisk is a Canadian program that uses accumulated evidence and experience to advise women about their actual risk of using a particular medication or being exposed to a particular xenobiotic in a current or planned pregnancy.^227,228

Although most women are primarily concerned about the teratogenic effects of medications, in utero exposure to therapeutic medications has other pharmacologic effects on newborn infants, such as hyperbilirubinemia or withdrawal syndromes.^39,78

Estimates of substance use in pregnancy vary tremendously, depending on the geographic location, practice environment, patient population, and screening method. In the National Survey on Drug Use and Health, approximately 16% of pregnant women smoked cigarettes, 11% drank ethanol, 5% used marijuana, 0.25% used cocaine, and 0.1% used heroin.²⁸⁵ Women tend to decrease their exposure to xenobiotics after they know they are pregnant.^33,150,153

PLACENTAL REGULATION OF XENOBIOTIC TRANSFER TO FETUSES

With respect to the transfer of xenobiotics from mother to fetus, the placenta functions in a manner similar to other lipoprotein membranes. Most xenobiotics enter the fetal circulation by passive diffusion down a concentration gradient across the placental membranes. The characteristics of a substance that favor this passive diffusion are low molecular weight (LMW), lipid solubility, neutral polarity, and low protein binding.⁹ Polar molecules and ions are transported through interstitial pores.⁹

Xenobiotics with a MW greater than 1,000 Da do not diffuse passively across the placenta, and this characteristic is used to therapeutic advantage. For example, warfarin (MW, 1,000 Da) easily crosses the placenta and causes specific fetal malformations.³⁴² However, heparin (MW, 15,000 Da), which is too large to cross the placenta, is not teratogenic and, consequently, is the preferred anticoagulant during pregnancy. Most medications have MWs between 250 and 400 Da and easily cross the placenta. For example, thiopental is highly lipid soluble and crosses the placenta rapidly. Fetal plasma concentrations reach maternal concentrations within a few minutes. Neuromuscular blockers such as vecuronium are more polar and cross the placenta slowly.⁹

Although ionization is a limiting factor for diffusion, some highly charged molecules will diffuse across the placenta. Valproic acid (pK_a of 4.7) is nearly completely ionized at physiologic pH, yet there is rapid equilibration across the placental membrane. The small amount of xenobiotic that exists in the nonionized form rapidly crosses the placenta, and as the equilibrium is reestablished, an additional, small amount of nonionized xenobiotic becomes available for diffusion.⁹

Fetal blood pH changes during gestation. Embryonic intracellular pH is high relative to the intracellular pH of the pregnant woman. During this developmental stage, weak acids diffuse across the placenta to the embryo and remain there because of “ion trapping.” Many teratogens, such as valproic acid, trimethadione, phenytoin, thalidomide, warfarin, and isotretinoin, are weak acids. Although ion trapping does not explain the mechanism of teratogenesis, it explains how xenobiotics accumulate in an embryo. Late in gestation, the fetal blood is 0.10 to 0.15 pH units more acidic than the maternal blood; this pH differential permits weakly basic xenobiotics to concentrate in the fetus during this period.⁹

The relative concentrations of protein-binding sites in the pregnant woman and fetus also have an impact on the extent of xenobiotic transfer to the fetus. As maternal free fatty acid concentrations increase near term, these fatty acids displace xenobiotics such as valproic acid or diazepam from maternal protein-binding sites and make more free xenobiotic available for transfer to the fetus.⁹ Fetal albumin concentrations increase during gestation and exceed maternal albumin concentrations by term. Because the fetus does not have high concentrations of free fatty acids to compete for protein-binding sites, these sites are available for binding the xenobiotics. At birth, when neonatal free fatty acid concentrations increase 2- to 3-fold, they displace stored xenobiotic from the binding protein. In the cases of valproic acid and diazepam, the elevated concentrations of free xenobiotic have adverse effects on the newborn infant.^{114,151,236,265}

The placenta also affects xenobiotic presentation to the fetus by ion trapping and xenobiotic metabolism. The placenta blocks the transfer of some positively charged ions such as cadmium and mercury¹²⁵ and actually accumulates them. This barrier does not necessarily protect the fetus, however, because these metal ions interfere with normal placental function and lead to placental necrosis and subsequent fetal death.

Beyond diffusion, specific transporters are found on both the fetal and maternal sides of the placenta. Transporters from the ATP-binding class (ABC) and the solute carrier class (SLC) have been identified.⁴⁶ These transporters help to deliver maternal solutes to the fetus and help the fetus to eliminate metabolic byproducts. They are also involved in the delivery of xenobiotics to the fetus.

The placenta contains xenobiotic-metabolizing enzymes capable of performing both phase I and phase II reactions (Chaps. 11 and 21). CYP1A1 is the primary CYP enzyme found in the placenta.⁴⁶ However, the concentration of biotransforming enzymes in the placenta is significantly lower than that in the liver, and it is unlikely that the level of enzymatic activity is protective for the fetus.²³⁰ Moreover, the fetus is exposed to reactive intermediates that form during these processes.

Placental transfer of xenobiotics has a positive effect as it permits delivery of desired therapies to the fetus. For example, if a fetus is found to have a supraventricular tachycardia or atrial flutter, digoxin given to the mother treats the tachydysrhythmia in the fetus.^145,231-234

EFFECTS OF XENOBIOTICS ON THE DEVELOPING ORGANISM

A basic premise of teratogenicity is that the particular toxic effects of a xenobiotic are determined by the stage of development of the embryo or fetus.³⁸ Although the fertilized ovum is generally thought to be resistant to toxic insult before implantation,³⁷ xenobiotics in the fallopian or uterine secretions prevent implantation of the embryo. Xenobiotic exposure leading to cell loss or chromosomal abnormalities also leads to loss of the embryo, possibly even before pregnancy is detected. If the preimplantation embryo survives a xenobiotic exposure, the functional cells usually proceed to normal development.²⁹⁸ Teratogens that act in such a manner elicit an “all-or-none response”; that is, the exposed embryo will either die or go on to normal development.

The dose–response curve of an environmental xenobiotic demonstrates deterministic (threshold) or stochastic (no threshold) effects.³⁸ Mutagenic and carcinogenic events are stochastic phenomena. Teratogenesis is a deterministic phenomenon with a threshold dose below which no effects occur. As the dose of the teratogen increases above the threshold, the magnitude of the effect increases. The effects might be the number of offspring that die or develop malformations or the extent or severity of malformations. Radiation has both deterministic effects (eg, microcephaly and growth retardation) and stochastic effects (induction of leukemia). Strictly speaking, teratogenic effects are those that occur at doses that do not cause maternal toxicity because maternal toxicity itself might be responsible for an observed adverse or teratogenic effect on the developing organism.³⁷

Organogenesis occurs during the embryonic stage of development between days 18 and 60 of gestation. Most gross malformations are determined before day 36, although genitourinary and craniofacial anomalies occur later.³⁷ The period of susceptibility to teratogenic effects varies for each organ system (Fig. 30–1). For instance, the palate has a very short period of sensitivity, lasting approximately 3 weeks, whereas the complete development of the central nervous system (CNS), including neurogenesis and differentiation, arborization, synaptogenesis and synaptic organization, and myelinization and gliogenesis, remains susceptible throughout the fetal period and into the neonatal period and infancy.

Theoretically, knowing the exact time of teratogen exposure during gestation would allow prediction of a teratogenic effect; this is true in animal models, where the dose and time of exposure can be strictly controlled. It is also true for thalidomide in which different limb anomalies are specifically related to exposures on particular days of gestation.³³⁷ In many clinical situations, for xenobiotics administered either for a short course or chronically, relating teratogenicity to a particular xenobiotic exposure is difficult because the exact time of conception and the exact time of exposure are unknown. This is particularly true when the primary exposure precedes the identification of pregnancy but there is secondary or ongoing exposure during pregnancy as xenobiotics are redistributed from tissue storage sites.¹⁷⁹

During the fetal period, formed organs continue their cellular differentiation and grow to functional maturity. Exposure to xenobiotics such as cigarettes and their toxic constituents during this period generally leads to growth retardation. Teratogenic malformations or death occur as a result of disruption or destruction of growing organs, as has resulted from exposure to angiotensin-converting enzyme inhibitors and angiotensin II, type I receptor blockers during the second and third trimesters.⁴⁵

Another concern during the fetal period is the initiation of carcinogenesis. Significant cellular replication and proliferation lead to a dramatic growth in size of the organism. At the same time, when the fetus is exposed to xenobiotics, development of biotransformation systems expose the organism to reactive metabolites that also initiate tumor formation. Some tumors, such as neuroblastoma, appear so early in postnatal life that their prenatal origin is likely. In pregnant rats given ethylnitrosourea during the embryonic period, lethal or teratogenic effects occur.²⁶⁷ If ethylnitrosourea is administered during the fetal period, there is an increased incidence of tumors in the offspring. Clear cell vaginal and cervical adenocarcinomas occur in the female offspring of women exposed to diethylstilbestrol (DES) during pregnancy.²⁶³

MECHANISMS OF TERATOGENESIS

Cytotoxicity is one mechanism of teratogenesis and is the characteristic result of exposure to alkylating or chemotherapeutic agents. Aminopterin, for example, inhibits dihydrofolate reductase activity and leads to suppression of mitosis and cell death. If exposure to a cytotoxic xenobiotic occurs very early in development, the conceptus can die, but sublethal exposure during organogenesis can result in maldevelopment of particular structures. There is evidence that after cell death, the remaining cells in an affected region try to repair the damage caused by the missing cellular elements. This “restorative growth” can lead to uncoordinated growth and exacerbate the original malformation.

In the case of cytotoxic agents, the mechanism of action is understood, although it is not always clear why particular xenobiotics affect particular structures. With other xenobiotics, the structural effects have a clearer relationship to the site of action. For instance, when corticosteroids are administered in large doses to some experimental animals during the period of organogenesis, malformations of the palate occur. Glucocorticoid receptors are found in high concentrations in the palate of the developing mouse embryo.²⁵⁹ Corticosteroid exposure had previously been thought to cause cleft palate in humans at a low frequency,^251,301 but recent work suggests there is no association between corticosteroid exposure and cleft palate.^25,307

Caloric deficiency is not considered teratogenic during the period of organogenesis; however, specific nutritional or vitamin deficiencies are. In particular, there is an increased incidence of neural tube defects (anencephaly and spina bifida) associated with dietary folate deficiency, although the specific mechanism of teratogenesis is unknown. To ensure that all women of childbearing age have adequate folate stores by the time they become pregnant and during their pregnancy and to reduce the number of severe birth defects, it is important for women of childbearing age to have folate supplementation either as vitamin or dietary supplements even before they become pregnant. Because fortification of grain products is considered critical to help accomplish this goal, the FDA requires US manufacturers to add folic acid to enriched breads, cereals, flours, corn meals, pastas, rice, and other grain products.²⁴⁴

Ethanol affects fetuses both directly and indirectly. The craniofacial malformations that occur in fetal alcohol syndrome (FAS) probably result from the effects of ethanol during the period of organogenesis. Growth retardation results from direct effects of ethanol on fetal growth or from indirect effects resulting from ethanol-induced maternal nutritional deficiencies.

MANAGEMENT OF ACUTE POISONING IN PREGNANT WOMEN

For most women, pregnancy and the postpartum period are a period of emotional happiness. However, women have a lifetime prevalence of depression that is 2 to 3 times higher than men, and for some women, psychiatric illness during pregnancy, particularly depression and anxiety, represents a significant comorbidity. A new first episode or recurrence of major depression occurs in approximately 3% to 5% of pregnant women and 1% to 6% of women in the postpartum period; an additional 5% to 6% of women during pregnancy and the postpartum period will have minor symptoms of depression.¹⁸⁸ Overall, these rates of depression are about the same in pregnant and nonpregnant women; however, during the postpartum period, the onset of new episodes of depression are higher than for nonpregnant women. Pregnant teenagers are also at higher risk of depression during pregnancy.¹⁹¹

Postpartum “baby blues” are common, typically involve relatively mild symptoms including mood swings, irritability, anxiety, and crying spells, and generally resolve by 2 weeks after delivery.¹⁹⁹ Postpartum blues do not include suicidal ideation. True depression is manifested by feelings of hopelessness or helplessness and include suicidal ideation. Postpartum psychosis, typically associated with bipolar disorder, is uncommon, but it represents a true psychiatric emergency because there is a high risk of suicide, infanticide, or both.²⁸

Risk factors for perinatal depression include an unplanned pregnancy, ambivalence about the pregnancy, poor social support, marital difficulties, adverse life events, and chronic stressors such as financial problems.^61,181,338 Of substantial importance is a personal or family history of depression, particularly previous perinatal depression. Discontinuation of antidepressant therapy represents a significant risk for relapse of disease, although relapse is possible even while a woman is receiving an antidepressant. Additional risk factors for depression include miscarriage, stillbirth, and preterm delivery.²⁹³

Depression in pregnancy has adverse effects on both the mother and the fetus.^32,316 For the pregnant woman, adverse effects include noncompliance with prenatal care; self-medication with tobacco, ethanol, and both licit and illicit xenobiotics drugs; poor sleep; poor appetite; and poor weight gain. In addition, there is an increased risk of spontaneous abortion, preeclampsia, preterm delivery, and fetal growth retardation. For many of these outcomes, it is difficult to differentiate the contributions from multiple confounding and interacting factors, including psychopathology, socioeconomic status, acute and chronic stressors, smoking, and the use of ethanol and other drugs of abuse.

One of the most extreme outcomes of depression is suicide. Suicidal ideation occurs in about 5% of pregnant women in community samples and up to 20% in women with underlying psychiatric illness.^111,191,240 Even so, suicide and suicide attempts during pregnancy are uncommon. Each year a small number of women die during pregnancy or the postpartum period; 1% to 5% of these pregnancy-related deaths are the result of suicide.^51,72,245 Between 2% and 12% of women who attempt or commit suicide are pregnant.^161,164,257 Overall, completed suicide occurs less frequently during pregnancy.^14,191,212 Psychiatric illness, including previous suicide attempts, predisposes to suicide attempts in pregnancy. Acute stressors leading to impulsive acts account for most of the uncompleted suicide attempts; these reasons include loss of a lover, economic crisis, prior loss of children, an unwanted pregnancy, and desire for an abortion.^68,187,350

Women who complete suicide typically have more severe psychiatric illness. In particular, these women are likely to use more violent means of suicide such as hanging or self-inflicted gunshot wounds, although poisonings are a significant contributor to these deaths.^90,245 In addition, substance and ethanol use are significant contributing factors in many cases; additionally, some pregnancy-related deaths are secondary to complications of substance use, such as overdose.²⁴⁵ Initiation of child protection proceedings, particularly in the setting of maternal substance use, is an additional risk factor for postpartum suicide.²⁴⁵ Some suicides follow pregnancy-related or neonatal complications.⁹⁰

Ingestion of xenobiotics is the most common method of attempted suicide during pregnancy and the postpartum period.^{61,109,257,261} There are approximately 7,000 xenobiotic exposures in pregnant women reported to the AAPCC annually, which account for approximately 2% of the exposures in adolescent and adult females of childbearing age (Chap. 130). The patterns of xenobiotic exposure are similar to those of adolescent and adult exposures in general with analgesics being the most common exposure. A similar pattern has also been reported for a California sample.²¹⁸ Some of these xenobiotic exposures are attempts to terminate pregnancy (Chap. 19).²⁵⁷ As with most poisonings overall, the severity of poisoning in pregnant women is typically minor.^240,245

In one US national sample, 1,659 pregnant women had poisoning-related hospitalizations, representing 0.04% of all hospitalized pregnant women and approximately 10% of injury-related admissions of pregnant women. A total of 244 of these women (15%) delivered their babies during these poisoning-related hospitalizations.¹⁷³ In a Swedish cohort, there was a higher rate of miscarriage than expected.¹⁰²

In national samples from both the United States and England, approximately 2% to 3% of deaths in pregnancy and the immediate postpartum period follow suicidal poisoning.^51,245 The AAPCC reports approximately 2 deaths per year in pregnant women, which represents approximately 2 per 1,000 AAPCC-reported adolescent and adult deaths overall, 4 per 1,000 deaths in adolescent and adult women, and approximately 3 deaths per 10,000 (0.03%) xenobiotic exposures in pregnancy (Chap. 130). In a combined Hungarian series of pregnant women who had suicidal poisonings, 19 of 1,044 (1.8%) died.⁶⁹

Any woman who attempts suicide during pregnancy or the postpartum period should have a psychiatric evaluation after she is medically stabilized. In particular, a growing number of specialized units or teams focus on pregnant and postpartum women and attempt to keep women and infants together in the postpartum period.^199,245

Managing any acute overdose during pregnancy provokes discussion of several questions. Is the general management different? Do altered metabolism and pharmacokinetics increase or decrease the woman’s risk of morbidity or mortality from a xenobiotic overdose? Is the fetus at risk of poisoning from a maternal overdose? Is there a teratogenic risk to the fetus from an acute overdose or poisoning? Is the use of an antidote contraindicated or should use be modified? When should a potentially viable fetus be emergently delivered to prevent toxicity? When should termination of a pregnancy be recommended?

As described earlier, physiologic changes during pregnancy affect pharmacokinetics; xenobiotics taken in overdose also have unpredictable toxicokinetics. In any significant overdose during pregnancy, pregnancy-related alterations in pharmacokinetics are unlikely to protect the woman from significant morbidity or mortality.

Although a single high-dose exposure to a xenobiotic during the period of organogenesis might seem analogous to an experimental model to induce teratogenesis, most xenobiotics ingested as a single, acute overdose do not induce physical deformities.⁷⁰ Antiepileptics are teratogenic and sometime are ingested in suicide gestures or attempts, but their teratogenicity is probably more directly related to chronic exposure. Acute acetaminophen (APAP) toxicity in the first trimester leads to an increased risk of spontaneous abortion,²⁷⁰ suggesting a teratogenic effect similar to the all-or-none response described earlier. In general, however, it is extremely difficult to ascribe teratogenicity to a particular xenobiotic exposure based on a single case report. There is, for example, a report of multiple severe congenital malformations in the stillborn fetus of a woman who overdosed on isoniazid during the 12th week of pregnancy.¹⁸⁵ However, because the background incidence of congenital malformations is 3% to 6%, it is almost impossible to determine for a single individual whether a particular exposure is the etiology of observed malformations.⁷¹ The Budapest Registry of Self-Poisoned Patients uses this construct to screen for possible teratogenic effects of xenobiotics in overdose.^69,70 Considering the successful outcome of most pregnancies that progress to term after an acute overdose, it is very unlikely that the small risk of teratogenesis would lead to a recommendation for termination of pregnancy after an acute overdose of most xenobiotics.⁸⁶

In general, any condition that leads to a severe metabolic derangement in a pregnant woman is likely to have an adverse impact on her developing fetus. Therefore, the management of overdose in a pregnant woman usually follows the principles outlined in Chap. 4, with close attention paid to the airway, oxygenation, and hemodynamic stability. The use of naloxone or dextrose has not been specifically assessed in pregnancy, but clinicians should be guided by the same considerations raised in managing the nonpregnant patient with alterations in respiratory or neurologic function. Opioid-induced respiratory failure in a pregnant patient will lead to fetal hypoxia and adverse effects; opioid withdrawal in a pregnant woman, whether induced by abstinence or the use of naloxone, can adversely affect the fetus or the pregnancy. Consideration of the benefits and risks of the use of naloxone for an opioid-poisoned woman in respiratory distress or coma suggests that reduced morbidity, for both the mother and fetus, is achieved by the use of carefully titrated doses of naloxone to minimize the likelihood of maternal withdrawal (Chap. 36 and Antidotes in Depth: A4).

Gastrointestinal decontamination is frequently a part of the early management of acute poisoning in nonpregnant patients. Gastric lavage is not specifically contraindicated for pregnant patients, and the usual concerns about airway management apply.

There is no specific contraindication to the use of activated charcoal in a pregnant woman. There is limited evidence for a role for whole-bowel irrigation (WBI) in the management of pregnant patients with certain xenobiotic exposures, for example, in the treatment of iron overdose in pregnancy.³³⁵ The use of oral polyethylene glycol-electrolyte solution (PEG) is safe in pregnant women.²³⁷

There have been no clinical trials on the use of antidotes in pregnant women in the context of acute poisoning. Deferoxamine is studied to treat chronic iron overload in pregnant women, and there is supportive data for this medication. Because there are no reports of adverse fetal effects from antidotal treatment of a poisoned pregnant woman, the benefit of antidotal therapy will almost invariably outweighs the risk of use in pregnancy. Conversely, in at least one case, withholding deferoxamine therapy contributed to the death of both a woman and her fetus.^207,320

ACETAMINOPHEN

Acetaminophen is the most common analgesic and antipyretic used during pregnancy and is also one of the most common xenobiotic overdoses during pregnancy. There are 3 published series of acute APAP overdose during all trimesters of pregnancy^220,221,270 in addition to multiple case reports. Overall, most pregnant women who continue their pregnancies recover from an APAP ingestion without adverse effects to themselves or their babies.

Twenty-eight women with first trimester exposures who continued their pregnancies are reported.^221,270 Nineteen women, 5 of whom had toxic serum APAP concentrations, delivered full-term infants. One woman with severe hepatotoxicity died along with the fetus. Eight women had spontaneous abortions; 5 of these 8 had toxic serum concentrations and received NAC (one within 8 hours and 4 between 12 and 17 hours after ingestion).

Thirty-two women with second-trimester exposures who continued their pregnancies are reported.^{105,197,221,270,275,317} Twenty-seven women, 7 of whom had toxic serum APAP concentrations, delivered full-term infants. One woman developed fulminant hepatic failure and underwent liver transplant at 20 weeks’ gestation; the infant developed severe complications in utero, and the woman terminated the pregnancy at 23 weeks. Two women, one with a nontoxic serum concentration and one who developed hepatotoxicity, delivered premature infants. Two women with nontoxic serum APAP concentrations had spontaneous abortions probably unrelated to the overdose.

Fifty-five women with third trimester exposures who continued their pregnancies are reported.^{20,47,66,107,126,141,172,184,221,256,270,274,279,281,287,322,340} Forty-two women, 15 of whom had toxic serum APAP concentrations, delivered full-term infants. Eleven women delivered premature infants. Five of them had toxic serum APAP concentrations and hepatotoxicity, 4 had toxic serum APAP concentrations without hepatotoxicity, and 2 had nontoxic serum APAP concentrations. Two women with severe hepatotoxicity delivered stillborn infants who also showed signs of hepatotoxicity and one woman with severe hepatotoxicity delivered an infant who died at 34 hours of life.

There are also several case reports of adverse pregnancy outcome in the setting of chronic use of APAP or acute overdose associated with chronic substance use.^{52,141,174,200,270,327} It is difficult to interpret these reports with respect to specific APAP effect because of the confounders of chronic disease, chronic APAP use, and the use of additional medications and substances.

Although APAP at recommended doses is considered safe in pregnancy,¹⁸⁹ in overdose, it puts the developing fetus at risk. As the cases demonstrate, APAP crosses the placenta to reach the fetus. The cases suggest an increased risk of spontaneous abortion after overdose during the first trimester, particularly in the setting of toxic serum APAP concentrations and delayed NAC therapy.²⁷⁰ The cases also suggest that overdose during the first trimester can lead to late sequelae, for instance, premature labor.

Some experimental work might explain early pregnancy loss after overdose. Acetaminophen prevented the development of preimplantation (2-cell stage) mouse embryos in culture, an effect that was not associated with alterations in glutathione concentrations,¹⁸² and led to abnormal neuropore development in cultured rat embryos.³¹⁵ These data suggest that APAP is directly toxic to the immature organism. However, other work reported that similar embryotoxic effects were associated with reductions in glutathione concentrations³⁴⁶ and that N-acetyl-p-benzoquinoneimine (NAPQI) produced nonspecific toxicity when added to the rat embryo culture medium.³¹⁵

The fetal liver has some ability to metabolize APAP to a reactive intermediate in vitro. Cytochrome P450 (CYP) activity was detected in intact hepatocytes, as well as in microsomal fractions isolated from the livers of fetuses aborted between 18 and 23 weeks of gestation. This fetal hepatic CYP activity increased between 18 and 23 weeks (the only period studied) but was maximally only 10% of the activity of hepatocytes isolated from adults after brain death.²⁷⁷ In 2 clinical cases, cysteine and mercapturate conjugates were identified in newborns exposed to APAP in utero, suggesting that the fetus and neonate metabolize APAP through the CYP system.¹⁸⁴ These data suggest that the fetus in utero and the neonate generate a toxic metabolite. The clinical cases suggest that the fetal liver is susceptible to injury, although whether this fetal hepatoxicity is related to fetal APAP toxicity, maternal toxicity, or postmortem changes is unclear.

This CYP activity has not been further characterized. However, CYP2E1, one of the cytochromes responsible for APAP metabolism, is present in human fetal tissues as early as 16 weeks of gestation.²²⁵ CYP3A4 and CYP1A2 are also involved in APAP metabolism but are not present in fetal liver. CYP3A7 is a functional fetal form of the CYP3 family, but its metabolic activity with respect to APAP is unstudied.¹²⁷

The clinical cases help address several questions regarding management of overdose during the third trimester (Table 30–2). Eight women whose pregnancies were all at less than 36 weeks of gestation developed hepatotoxicity. Two infants died in utero and one infant died on the second day of life with evidence of hepatotoxicity. The other 5 infants experienced problems associated with prematurity but did not develop obvious hepatotoxicity. One of these 5 infants had 4 exchange transfusions to treat a cord acetaminophen concentration of 76 mcg/mL and had an unexplained death at 3 months of age. Five women whose pregnancies were all at 36 or more weeks of gestation did not develop hepatotoxicity. One infant had 2 exchange transfusions to treat a cord acetaminophen concentration of 217 mcg/mL and did not develop hepatotoxicity but died of sudden infant death syndrome (SIDS) at 5 months of age. One infant received intravenous (IV) NAC and had a transient elevation of aspartate aminotransferase (AST) and prothrombin time. Two infants were not treated, and both did well, although one had a transient elevation of AST. One infant was born 6 weeks after the overdose and was normal.

Taken together these cases suggest that when there is maternal hepatotoxicity there is a moderate risk of fetal hepatotoxicity and fetal morbidity. There is also a moderate risk of prematurity delivery.

Urgent delivery is often recommended in the setting of severe maternal hepatoxicity that is associated with fetal distress. Although a fetus with prolonged exposure to APAP in utero is at risk of developing severe hepatotoxicity, not all at-risk infants have been affected. The contribution of gestational age, maternal disease state, or other maternal factors to fetal or neonatal toxicity is unknown. Although there are insufficient case data to suggest that APAP overdose per se is an indication for urgent delivery, some authorities recommend urgent delivery when the maternal serum APAP concentration is in the toxic range but hepatotoxicity has not yet developed.³²⁴ The clinical cases also suggest that significant APAP overdose with or without associated hepatotoxicity can precipitate premature labor and that even women with nontoxic serum concentrations are at an increased risk.

In 2 cases, exchange transfusion was used to treat the exposed neonate. In both cases, the APAP half-life was prolonged, and in neither case was this affected by the transfusion. Disturbingly, these 2 infants had unexplained deaths at several months of age. There are currently no data to support the use of exchange transfusion as therapy for elevated serum APAP concentrations in a neonate.

A pregnant woman with acute toxic APAP ingestion should be treated with NAC (Chap. 33 and Antidotes in Depth: A3). This therapy is intended to treat the mother. Maternal hepatoxicity and delayed NAC therapy are associated with fetal toxicity;²⁷⁰ whether NAC therapy can prevent later sequelae such as premature delivery is unclear. Although NAC did not cross the sheep placenta in vivo²⁹⁹ and the perfused human placenta in vitro, NAC was found in cord blood after it was administered to 4 mothers before delivery.¹⁴¹ Even if NAC does cross the placenta, whether it can prevent fetal hepatotoxicity is unknown because not all exposed fetuses develop hepatotoxicity.

Use of APAP during pregnancy may be associated with other risks. One area of concern is the association of prenatal exposure to APAP with the development of childhood asthma. Several meta-analyses have reviewed many of the published studies.^57,91,93 Many of those studies were criticized for lack of control of potential confounders such as the stage of pregnancy, dose and frequency of use, variable definitions of asthma, and the different childhood age groups considered. One of the biggest concerns is the possible confounding by indication because why a pregnant woman took acetaminophen, whether for a respiratory tract infection or for back pain for example, may be important as a predictor. Notwithstanding these concerns, the meta-analyses report a positive association between prenatal exposure to APAP and childhood wheezing and/or asthma with an OR of approximately 1.28-1.5.

IRON

Intentional iron exposures during pregnancy are reported; maternal toxicity is generally greater than fetal toxicity. In 2 cases, normal babies were delivered, although the mothers died.^247,268 In another case, the mother had severe iron toxicity with metabolic acidosis, shock, kidney failure, and disseminated intravascular coagulation but was not treated with deferoxamine because of concerns about its teratogenic risks. Instead, the mother received an exchange transfusion followed 45 minutes later by a spontaneous abortion of the 16-week fetus.^207,320 Neonatal and cord blood iron concentrations were not elevated. In several cases, pregnant women who had signs and symptoms of iron poisoning and elevated serum iron concentrations were treated with deferoxamine and subsequently delivered normal babies.^{30,162,177,262,292,330}

Although the placenta transports iron to the fetus efficiently,¹⁹ it also blocks the transfer of large quantities of iron. In a sheep model of iron poisoning, only a small amount of iron was transferred across the placenta despite significantly elevated serum iron concentrations.⁶⁷

Deferoxamine is an effective antidote for iron poisoning (Chap. 45 and Antidotes in Depth: A7), but it is reported to be a teratogen that causes skeletal deformities and abnormalities of ossification when given to pregnant rabbits at high doses. A rat model observed similar effects but only with doses of deferoxamine that caused maternal toxicity.³⁴ Experimentally, in sheep, deferoxamine was minimally transported across the placenta;⁶⁷ therefore, the reported fetal effects could be secondary to chelation of essential nutrients (eg, trace metals) on the maternal side of the placenta.³²⁴

In clinical case reports of iron overdose for which deferoxamine was used, there have been no adverse effects on the fetus, although most have been either second or third trimester poisonings.^{30,162,177,247,262,292,330} In a case series of 49 patients with iron poisoning during pregnancy, few of the patients exhibited any clinical toxicity other than vomiting and diarrhea; 25 received deferoxamine, most by the oral route.²¹⁹ One woman with a first trimester overdose, 8 women with second trimester overdoses, and 12 women with third trimester overdoses were treated with deferoxamine and subsequently delivered full-term infants. One infant whose mother overdosed at 30 weeks of gestation had webbed fingers on one hand. One woman who overdosed at 20 weeks had minimal clinical toxicity, received deferoxamine, and delivered a 2.5-kg male infant at 34 weeks of gestation. One woman with a first trimester overdose and 2 women with second trimester overdoses elected to terminate their pregnancies.

Further support for the safe use of deferoxamine in pregnancy is the experience with its use for pregnant women with thalassemia major. For many years, deferoxamine has been administered as part of the therapy for posttransfusion iron overload without adverse effects.³⁰⁵ Considering the potentially fatal nature of severe iron poisoning, deferoxamine should be administered when signs and symptoms indicate significant poisoning.

Iron overdose is one of the few specific indications for WBI, because iron is not adsorbed to activated charcoal (Antidotes in Depth: A1 and A2). A case report demonstrated elimination of pill fragments after treatment of a pregnant woman with WBI.³³⁵

CARBON MONOXIDE

Carbon monoxide is the leading cause of unintentional poisoning fatalities in the United States. In contrast to iron and most other xenobiotics, when pregnant women are exposed to carbon monoxide, the fetus is at greater risk of toxicity than the mother. There are reports of both the mother and fetus dying; the mother surviving but the fetus dying; and both the mother and fetus surviving but with an adverse neonatal outcome, primarily brain damage resembling that seen after severe cerebral ischemia.^{48,65,169,194,208,242,335,353} Similar clinical effects are also observed in animal models.^83,115,195

The case literature suggests an increased risk of poor fetal outcome with clinically severe maternal poisoning or significantly elevated carboxyhemoglobin concentrations.^169,242 Women with minimal symptoms or low concentrations of carboxyhemoglobin have a low risk of fetal toxicity, but a lower limit of exposure without effect has not been specifically defined.¹⁶⁹

In animal models, under experimental conditions, the fetus has a carboxyhemoglobin concentration 10% to 15% higher than the mother.^135,195 After exposure to carbon monoxide, the fetus achieves peak carboxyhemoglobin concentrations 58% higher than those achieved by the mother at steady state, and the time to peak concentration is also delayed compared with that of the mother. Similarly, the elimination of carbon monoxide occurs more slowly in the fetus than in the mother.^135,194,195 One case report describes such a phenomenon: after one hour of supplemental oxygen, the maternal carboxyhemoglobin concentration was 7% and the fetal carboxyhemoglobin concentration was 61% at the time of death in utero.⁸⁸

Carbon monoxide leads to fetal hypoxia by several mechanisms: (1) maternal carboxyhemoglobin leads to a decrease in the oxygen content of maternal blood, and therefore less oxygen is delivered across the placenta to the fetus, which normally has an arterial PO₂ of only 20 to 30 mm Hg; (2) fetal carboxyhemoglobin causes a decrease in fetal PO₂; (3) carbon monoxide shifts the oxyhemoglobin dissociation curve to the left and decreases the release of oxygen to the fetal tissues (an exacerbation of the physiologic left shift found with normal fetal hemoglobin); and (4) carbon monoxide inhibits cytochrome oxidase or other mitochondrial functions.

The treatment for severe carbon monoxide poisoning is hyperbaric oxygen therapy (HBO) (Chap. 122 and Antidotes in Depth: A40). There are questions about the use of HBO in pregnant women because animal models suggest HBO adversely affects the embryo or fetus.^{98,226,290,323} The applicability of these animal models to humans is difficult to assess; many of the animal models used hyperbaric conditions of greater pressures and duration than those clinically used for humans.

Hyperbaric oxygen therapy has been used therapeutically for carbon monoxide poisoning in pregnancy with good results reported, although there are limited data on the long-term follow-up of the children.^{41,97,108,120,136,169,336} One large series reported 44 women who were exposed to carbon monoxide during pregnancy and were treated with HBO, regardless of clinical severity or gestational age: 33 had term births, one had a premature delivery 22 weeks after HBO during an episode of maternal fever, 2 had spontaneous miscarriages (one 12 hours after severe poisoning and one 15 days after mild poisoning), one delivered a child with Down syndrome, one had an elective abortion, and 6 were lost to follow-up.⁸⁸ Details regarding trimester of exposure, maternal carboxyhemoglobin concentration, and severity of symptoms are not available, making it difficult to interpret the reported adverse outcomes. Although HBO appears to be safe for pregnant women and seems to present little risk to the fetus, it is not clear whether HBO can prevent carbon monoxide–related fetal toxicity.

Hyperbaric oxygen is appropriate therapy for any pregnant woman exposed to carbon monoxide with an elevated serum carboxyhemoglobin concentration or evidence of fetal distress. To allow the fetal carboxyhemoglobin to be eliminated, if HBO therapy is not available, 100% oxygen is recommended for all pregnant women for 5 times as long as the time needed for her carboxyhemoglobin concentration to return to the normal range.¹⁹⁴ Thus, if a pregnant woman’s carboxyhemoglobin concentration returns to normal in one hour, she should continue to receive 100% oxygen for at least 5 hours.

SUBSTANCE USE DURING PREGNANCY

Substance use during pregnancy leads to complex medical and psychosocial effects for the pregnant woman as well as the fetus and newborn infant. With the increased use of cocaine during the latter half of the 1980s and 1990s, there was great interest in determining the effects of cocaine use during pregnancy. As research in this area progressed, many of the critical methodologic issues related to substance use research were highlighted.^{99,149,186,238,354}

Substance-using women often have multiple risk factors for adverse pregnancy outcomes, such as low socioeconomic status, polysubstance use, ethanol and cigarette use, sexually transmitted infections, AIDS, malnutrition, and lack of prenatal care. Lack of prenatal care is highly correlated with premature birth, and smoking is associated with spontaneous abortion, growth retardation, and SIDS.^158,348 Other factors not specifically related to substance use such as age, race, gravidity, and prior pregnancies also affect pregnancy outcome. Each of these factors represents a significant potential confounding variable when the effects of a particular xenobiotic such as ethanol or opioids are evaluated during pregnancy and must be controlled for in research design. Many of these factors are also significant confounders in evaluation of postnatal growth and development.

Bias can occur in the selection of study subjects. For example, if all the patients are selected from an inner-city public hospital obstetrics service, minorities may be over-represented. On the other hand, if all the patients are selected from substance-using populations of rural Vermont or West Virginia, minorities may be under-represented. There may be significant differences in socioeconomic status or genetics. If cohorts are followed over a long time, study subjects are frequently lost to follow-up. The ones who continue may be more motivated or may have more problems that need attention?

Categorizing patients into substance-use groups is difficult. Self-reporting of substance use is frequently unreliable or inaccurate and making determinations about the nature, frequency, quantity (dose), or timing (with respect to gestation) of xenobiotic exposure is difficult. Because substance users frequently use multiple xenobiotics, it is difficult to categorize subjects into particular xenobiotic-use groups, and patients using different xenobiotics can be inadvertently grouped together. In fact, there probably are no actual xenobiotic-free control groups.

When urine drug screens are used to identify substance users, there is a high probability of false negatives because drug screens reflect only recent use. This factor is particularly important because substance use tends to decrease later in pregnancy and a negative urine drug screen in the third trimester or at delivery will fail to identify a woman who was using xenobiotics early in pregnancy. Testing for xenobiotics in hair or meconium can improve the accuracy of the analysis with regard to the entire pregnancy.^35,167

Another bias involves selection of infants who are exposed to xenobiotics. Evaluating newborns who are “at risk,” show signs of substance withdrawal, or have positive urine drug screens will miss some exposed infants. When research concerns the neurobehavioral development of children exposed in utero to substances, it is important that the examiners performing the evaluation be blinded to the infants’ xenobiotic exposure category. Finally, there is bias against publishing research that shows a negative or nonsignificant effect.¹⁶⁸

Ethanol

Approximately 9% of pregnant women (200,000) 15 to 44 years of age report consuming ethanol during the previous month, 4.6% (100,000) report binge drinking, and 0.8% (20,000) report heavy drinking. Of nonpregnant women 15 to 44 years of age, about 55% report consuming ethanol during the previous month, 30% report binge drinking, and 6% report heavy drinking.²⁸⁶ The highest prevalence of drinking is during the first trimester and decreases in the second and third trimesters. These numbers are consistent with the CDC estimate of a 7.3% risk of an ethanol-exposed pregnancy.¹¹⁹

Chronic ethanol use during pregnancy produces a constellation of fetal effects. Fetal alcohol spectrum disorders (FASD) is an umbrella term describing the range of effects that can occur in an individual whose mother drank ethanol during pregnancy.^29,311

The most severe effects occur in the FAS which is characterized by (1) intrauterine or postnatal growth retardation; (2) intellectual disability or behavioral abnormalities; and (3) facial dysmorphogenesis, particularly microcephaly, short palpebral fissures, epicanthal folds, maxillary hypoplasia, cleft palate, a hypoplastic philtrum, and micrognathia.^29,206 A child can be diagnosed with FAS even when a history of regular gestational ethanol use cannot be confirmed.

In an attempt to formalize diagnostic criteria for FAS and other gestational ethanol-related effects, the Institute of Medicine proposed some additional descriptors that are in common use and which have been refined.^143,318 Partial FAS is applied to a child with some of the characteristic facial features and with growth retardation, neurodevelopmental abnormalities, or other behavioral problems. Alcohol-related birth defects (ARBD) are congenital anomalies other than the characteristic facial features described earlier, for example, cleft palate, which sometimes occurs with regular gestational ethanol use. Alcohol-related neurodevelopmental disorder (ARND) describes neurodevelopmental abnormalities or other behavioral problems, which sometimes occur with regular gestational alcohol use.

Differential expression of the syndrome reflects the effects of varying quantities of ethanol ingested at critical periods specific for particular effects. The craniofacial anomalies probably represent teratogenic effects during organogenesis but some CNS abnormalities and growth retardation result from adverse effects later in gestation.

There is considerable controversy about what amount of ethanol consumption causes FASD.^235,243 Some researchers believe that FASD is a result of alcoholism—chronic regular use or frequent binging—rather than to low concentrations of gestational ethanol exposure, no matter how little or how infrequent.^2,118 A high level of consumption would be in the range of 50 to 100 mL of 100% ethanol (4–6 “standard” drinks of hard liquor) regularly throughout pregnancy³¹⁸ or binge drinking (at least 5 standard drinks per occasion), with a significantly elevated peak blood ethanol concentration.^101,215 However there are reports of behavioral abnormalities associated with the consumption of as little as one drink per week.³¹² In this regard, neither a no-effect amount nor a safe amount of ethanol use in pregnancy has been determined,¹⁵⁶ and therefore the US Surgeon General recommends no ethanol consumption during pregnancy.²⁴⁶

The incidence of FAS in the United States is reported as 0.3 to 0.8 per 1,000 children when measured by passive surveillance and record review.¹⁰³ In-person expert assessment of a community-based school-age sample reported an incidence of 6 to 9 per 1,000 children.²¹⁴ This means that several hundred children with FAS and several thousand with fetal ethanol effects will be born each year; thus, ethanol use is considered the leading preventable cause of intellectual disability in this country.³¹⁸ Although the primary determinant of FAS and its effects is the amount of maternal ethanol consumption, there is some evidence that paternal ethanol exposure plays a contributing role.¹

Ethanol use during pregnancy leads to an increased incidence of spontaneous abortion, premature deliveries, stillbirths,²⁰⁵ neonatal ethanol withdrawal,^26,241 and possibly carcinogenesis.¹⁶³ Infants can be irritable or hypertonic and can have problems with habituation and arousal. Long-term behavioral and intellectual effects include decreased IQ, learning disabilities, memory deficits, speech and language disorders, hyperactivity, and dysfunctional behavior in school.^213,312,319

Autopsies of children with FAS demonstrate malformations of gray and white matter, a failure of certain regions such as the corpus callosum to develop, a failure of certain cells such as the cerebellar astrocytes to migrate, and a tendency for tissue in certain regions to die.⁹⁶ The mechanisms of ethanol-induced teratogenesis are not fully elucidated.^27,128 Much of the work in animals has focused on the developing nervous system, where ethanol adversely affects nerve cell growth, differentiation, and migration, particularly in areas of the neocortex, hippocampus, sensory nucleus, and cerebellum.^123,321

Several mechanisms are potential contributors to the effects of ethanol.^117,222 Ethanol interferes with a number of different growth factors that affect neuronal migration and development.³⁴⁹ In addition, ethanol interferes with the development and function of both serotonin and N-methyl-D-aspartate (NMDA) receptors. Ethanol, or possibly its metabolite acetaldehyde, also causes cell necrosis directly or through the generation of free radicals and excessive apoptosis.^60,96,132 In particular, craniofacial abnormalities could be related to apoptosis of neural crest cells through the formation of free radicals, a deficiency of retinoic acid, or the altered expression of homeobox genes that regulate growth and development.³²¹

One integrative model of ethanol-induced teratogenesis proposes that sociobehavioral risk factors, such as drinking behavior, smoking behavior, low socioeconomic status, and cultural or ethnic influences, create provocative biologic conditions, such as high peak blood ethanol concentrations, circulating tobacco constituents, and undernutrition. These provocative factors exacerbate fetal vulnerability to potential teratogenic mechanisms, such as ethanol-related hypoxia or free radical–induced cell damage.³

Opioids

Opioid use in pregnancy remains a significant cause of both maternal and neonatal morbidity. Over the past 2 decades, use and misuse of opioids, particularly prescription opioids, has been spiraling upward in the United States. Of women 15 to 44 years of age, 25% to 40% are prescribed opioids, in particular hydrocodone, codeine, and oxycodone.⁷ These prescription rates are higher for women insured by Medicaid compared to private insurance and for women living in the south compared to the northeast. Similarly, opioid prescription rates for pregnant women are also high, ranging from 10% in a Veterans Administration sample, to 14% in a commercial insurance sample, to 29% in a Medicaid sample.^23,76,89,171 Again there are significant regional (state to state) variations as well as urban versus rural differences in these prescribing patterns during pregnancy that reflect general trends in opioid prescribing.^76,339 Hydrocodone, codeine, and oxycodone are also the primary opioids prescribed during pregnancy.

Past month surveys provide a different perspective. Approximately 21,000 pregnant women 15 to 44 years of age report illicit and prescription opioid misuse and abuse (approximately 0.2% of pregnant women report heroin use, 0.8% report the illicit use of prescription analgesics, and approximately 0.1% report the use of oxycodone).^284,286,309

Pregnant opioid users are at increased risk for many medical complications of pregnancy, such as hepatitis, sepsis, endocarditis, sexually transmitted infections, and AIDS, and at increased risk for obstetric complications, such as miscarriage, premature delivery, or stillbirth.^99,116 Some of the obstetric complications are related to associated risk factors in addition to the opioid use. Maternal opioid use most commonly affects fetal growth.^99,116,354 There is an increased incidence of low birth weight in babies born to opioid-using mothers compared with control participants, and the effect is greater for heroin than for methadone.

There is extensive clinical experience with Medication Assisted Therapy (MAT) with methadone maintenance during pregnancy. Compared with untreated heroin use, methadone maintenance during pregnancy is associated with better obstetric care, decreased risk of maternal withdrawal, increased fetal growth, decreased fetal mortality, decreased risk of HIV infection, and increased likelihood that a child will be discharged home with his or her parents.¹⁵⁴ Women who receive low-dose methadone and good prenatal care are at increased risk for pregnancy-related complications but have birth outcomes similar to nonusers.⁹⁹

Just as alternative opioid replacement strategies with buprenorphine were introduced for opioid users in general, buprenorphine is more frequently being used during pregnancy especially for women maintained on buprenorphine before becoming pregnant. In one large trial, obstetric outcomes were similar after buprenorphine treatment compared with methadone although more women in the buprenorphine arm dropped out of the study because they did not feel well during the induction of therapy.^154,155 Similar results regarding obstetric outcomes were reported in one study primarily looking at neonatal abstinence syndrome (NAS).⁴² In another large cohort of women on MAT, a small number of women switched from buprenorphine to methadone for a variety of reasons, but some birth outcomes such as gestational age and birth weight were a little better in the buprenorphine group.²²⁴ One group in Australia reported the use of implantable naltrexone in pregnant patients after opioid detoxification.^146,147

The most significant acute neonatal complication of opioid use during pregnancy is the NAS, which is characterized by hyperirritability; GI dysfunction; respiratory distress; and vague autonomic effects, including yawning, sneezing, mottling, and fever.¹⁴⁴ Myoclonic jerks or seizures also signify neurologic irritability. Withdrawing infants are recognizable by their extreme jitteriness despite efforts at consolation. Approximately 50% to 90% of chronic opioid-exposed newborns show some signs of NAS.^99,154

Just as opioid use and misuse has increased among the general population overall and pregnant women in particular, so has the frequency of NAS increased in the United States and abroad. In the United States, between 2000 and 2012 the rate of NAS increased from 1.2 per 1,000 to 5.8 per 1,000 live births.^253,255 Again there is significant regional variation. The rate of NAS varied from a low of 0.7 per 1,000 births in Hawaii to 33.4 per 1,000 in West Virginia¹⁶⁵ and is more prevalent in rural compared to urban areas.³³⁹ In 2012, the average cost of hospitalization was $66,700 for a baby with NAS and $93,400 for a baby with NAS treated pharmacologically as compared to $3500 for an uncomplicated term newborn. Considering that approximately 4% of neonatal ICU days overall and in some centers as much as 40% to 50% of neonatal ICU days are attributed to babies with NAS, much of this cost is related to an increased number of infants admitted to NICUs for therapy and to prolonged lengths of stay for treatment.³²⁹ Approximately 80% of the 2012 aggregate cost of 1.5 billion dollars was covered by Medicaid.

The Neonatal alcohol abstinence syndrome typically occurs within one week of birth. Heroin withdrawal usually occurs within the first 24 hours whereas methadone and buprenorphine withdrawal usually occur within 2 to 3 days of delivery. The onset of methadone withdrawal is delayed compared to heroin because methadone has a larger volume of distribution and slower metabolism in the neonate and therefore an increased half-life. The onset and severity of methadone withdrawal may be related to the absolute neonatal serum concentration or possibly to the rate of decline in concentration.^80,175 In one study, methadone withdrawal occurred when the plasma concentration fell below 0.06 mg/mL.²⁷⁸

As mentioned previously, the onset and severity of withdrawal is related to which opioids were used; which other prescription medications or substances were used; how much of these prescription medications or other substances was used long-term; how much was used near the time of delivery; the character of the labor; whether analgesics or anesthetics were used; and the maturity, nutrition, and medical condition of the neonate.⁷⁷ Acute NAS generally last from days to weeks but as many as 80% of infants were reported to have recurrence of some symptoms such as restlessness, agitation, tremors, wakefulness, hyperphagia, colic, or vomiting for 3 to 6 months.⁵⁴ In general, the incidence and severity of withdrawal are greater from methadone than from either heroin or buprenorphine.¹⁴⁴

In general, babies who are extremely irritable; have feeding difficulties, diarrhea, or significant tremors; or cry continuously should receive pharmacological therapy.¹⁴⁴ The severity of withdrawal as measured by a standardized scoring scale is typically used to determine the need for therapy; several scales are in use although none has been validated.^99,192 Although most hospitals have treatment protocols for NAS, there is significant variation in their use.^31,254 Reducing this variation by training staff to apply these scores in a consistent manner leads to reductions in length of treatment, length of hospital stay, and total dose of opioid administered.¹⁸

Severe NAS often require pharmacologic therapy, and the typical approach was to provide care in a NICU. However, considering that treatment of withdrawal is recommended to begin with swaddling or tightly wrapping the infant, minimal handling or stimulation, and demand feeding, all in a quiet and comforting environment, the NICU is increasingly being seen as a less-than-ideal environment to provide care. In fact, this phase of treatment is generally neither standardized nor controlled in clinical trials of pharmacologic therapy.¹²²

In one center where a standardized approach to treatment based on a rooming-in model with increased family involvement was implemented on an inpatient non-NICU ward, the proportion of infants treated with morphine decreased from 46% to 27% and the morphine dose decreased from 13.7 mg to 6.6 mg per infant.¹³⁷ In addition, the length of stay (LOS) decreased from 16.9 to 12.3 days and the average cost decreased from $19,737 to $8,755 per treated infant. This center also used a standardized treatment and weaning protocol. In another center using a similar approach, the proportion of infants treated decreased from 98% to 14%, the average LOS decreased from 22.4 days to 5.9 days, and the average cost decreased from $44,824 to $10,259 per infant.¹²¹ In contrast to almost every other study that uses the original or modified version of the Finnegan neonatal abstinence scoring system to determine the need for pharmacologic therapy, this group developed a score focused only on what might be considered an infant’s primary needs—the ability to eat, sleep, and be consoled. Pharmacologic therapy was instituted only if maximal nonpharmacologic intervention was unsuccessful and these primary needs could not be fulfilled.

Opioid agonists such as morphine, methadone, buprenorphine, tincture of opium, and paregoric; sedative–hypnotics such as diazepam and phenobarbital; and clonidine have all been used both individually and in various combinations to treat withdrawal. Most centers primarily use oral morphine to treat withdrawal because it is a short-acting pure opioid agonist and the formulation has no additives.²⁶⁰ Buprenorphine is being increasingly explored as a primary treatment option. Clonidine and phenobarbital are typically used as adjunctive therapies. Few well-controlled trials have evaluated the relative efficacy of various approaches or examined long-term effects of therapy.^248,249

At one center, sublingual buprenorphine was compared to oral morphine.^154,155 In these trials, buprenorphine was associated with a 5- to 15-day (18%–46%) decreased length of opioid treatment (LOT) and an 11- to 19-day (29%–36%) decreased length of stay (LOS). In one trial comparing buprenorphine to methadone, both the average LOT and the average LOS were both approximately 4.5 days less in the buprenorphine group. In this study, the LOT for buprenorphine was shorter than the LOT in 2 of the 3 buprenorphine versus morphine studies. It is hard to know exactly how to compare these findings, considering that the studies have different patient populations and different treatment algorithms, and that regardless of the opioid chosen the LOT and LOS can decrease in association with standardized opioid treatment and weaning protocols as well as with a standardized and intensified approach to nonpharmacologic intervention. Nonetheless, the evidence suggests that the LOT and LOS are both shorter after buprenorphine compared to either morphine or methadone.

Five percent to 7% of babies showing signs of withdrawal experience seizures, generally by 10 days after birth.¹³⁴ Seizures occur more frequently after methadone withdrawal than after heroin withdrawal. These seizures do not necessarily predispose to idiopathic epilepsy; in one small study, children who had withdrawal seizures were without seizures and had normal neurologic examination findings and psychometric testing at one-year follow-up.⁸¹

Opioid agonists are more effective at preventing withdrawal seizures from heroin or methadone than phenobarbital or diazepam are.^134,157 However, sedative–hypnotics are commonly abused by pregnant heroin users or pregnant women maintained on methadone. Therefore sedative–hypnotic withdrawal seizures can contribute to the overall neonatal abstinence symptomatology, and for this reason, many treatment protocols include the use phenobarbital as a treatment adjunct.

Infants of opioid-using mothers have a 2 to 3 times increased risk for SIDS compared with control participants.^158,159 Proposed mechanisms include a decreased medullary responsiveness to CO₂ or some condition related to the postnatal environment.⁹⁹

Although young children born to opioid users do not seem to have significant differences in behavior compared to control participants, older children have increased learning problems and school dysfunction particularly related to behavior difficulties.³⁵⁴

Cocaine

Approximately 0.1% of pregnant women in the United States between the ages of 15 and 44 years report previous-month use of cocaine; for 2016, this represented only 1,000 women.²⁸⁶ This is in contrast to approximately 431,000 female cocaine users between the ages of 15 and 44 years who reported previous-month use for the same period. Compared to the numbers reported during the cocaine epidemic of the 1990s, this is a significant reduction in exposure. Nonetheless, the number of children exposed to cocaine in utero over the past 30 years has been estimated at 7.5 million.⁵⁰ The consequences of cocaine use during pregnancy have been extensively reviewed.^{139,178,210,258}

The incidence of abruptio placentae may be significant when related to acute cocaine use.³⁰³ Some uncommon perinatal problems include seizures, cerebral infarctions, and intraventricular hemorrhage.^58,258

Cocaine use is significantly related to decreases in gestational age, birth weight, length, and head circumference, although these growth parameters generally correct by several years of age.^178,210 In addition, these growth effects are exacerbated by concomitant ethanol, tobacco, and opioid use.^178,210 Good prenatal care can mitigate many of these adverse effects of cocaine.^201,269,354

Significant congenital malformations were reported among some infants who were exposed to cocaine in utero, specifically genitourinary malformations, cardiovascular malformations, and limb-reduction defects.^44,106 However, in one large population-based study, there was no increase in the incidence of malformations.²⁰⁹

Teratogenic effects are observed in animal models of in utero cocaine exposure. Decreased maternal and fetal weight gain and an increased frequency of fetal resorption were demonstrated in rats⁹⁵; sporadic physical anomalies are also observed.⁵⁹ Teratogenic effects similar to those observed in humans were reported in mice, including bony defects of the skull, cryptorchidism, hydronephrosis, ileal atresia, cardiac defects, limb deformities, and eye abnormalities.^{100,202,203,223} Cocaine caused hemorrhage and edema of the extremities and, subsequently, limb-reduction defects in rats when administered during midgestation in the postorganogenic period.³⁴³

The perinatal effects of cocaine are probably mediated through a vascular mechanism. Cocaine administration in the pregnant ovine model causes increased uterine vascular resistance, decreased uterine blood flow, increased fetal heart rate and arterial blood pressure, and decreased fetal PO₂ and O₂ content.^17,351 Similar effects are reported in rats.²⁵² Fetal hypoxia causes rupture of fetal blood vessels and infarction in developing organ systems such as the genitourinary system^56,223,308 or the CNS.^53,79,344 Hyperthermia or direct effects of cocaine in the fetus could exacerbate these effects.³⁶ Limb-reduction defects similar to those attributed to cocaine are produced after mechanical clamping of the uterine vessels.^36,345 A developing concept is that after vasospasm and ischemia, reperfusion occurs with the generation of oxygen free radicals and subsequent injury.^351,352

Despite the reported malformations and a possible mechanism neither the human epidemiology nor the effects observed in animal models suggest a specific teratogenic syndrome. The risk of a significant malformation from prenatal cocaine exposure is low, but the effect, if one occurs, could be severe.^37,84,106

One of the greatest concerns about prenatal cocaine exposure is the potential adverse effect on the developing child, and this is an intensive area of epidemiologic research. The most common findings in early infancy are lability of behavior and autonomic regulation; decreased alertness and orientation; and abnormal reflexes, tone, and motor maturity. However, many studies show no effect, especially after controlling for confounding variables.^94,178,210 For some children, effects manifest in later infancy as difficulty with information processing and learning. For school-age children, observed cognitive deficits are often related to the home environment even for children who showed some of the typical neonatal behaviors.^104,178,210 Nonetheless, there is evidence of impairment in modulating attention and impulsivity, which makes handling unfamiliar, complex, and stressful tasks more difficult,^178,210,216 and these effects are also observed in animal models of prenatal cocaine exposure.^{84,129,190,313}

The mechanism of neurotoxicity is not specifically elucidated. As described earlier, for many of the maternal and fetal physical defects, cocaine has direct toxicity, or alternatively, effects could be mediated through hypoxia or oxygen free radicals. Because cocaine interferes with neurotransmitter reuptake, it is likely that cocaine also disrupts normal neural ontogeny by interfering with the trophic functions of neurotransmitters on the developing brain, in particular dysfunction of signal transduction via the dopamine D₁ receptor.^129,190,217 These mechanisms are important in the etiology of neurobehavioral effects.

BREASTFEEDING

In the United States, breastfeeding is the recommended method of infant nutrition because it offers nutritional, immunologic, and psychological benefits without additional costs. Many women use prescription and nonprescription xenobiotics while breastfeeding and are concerned about the possible ill effects on their infants of these xenobiotics in the breast milk. This concern extends to the possible exposure of infants to occupational and environmental xenobiotics via breast milk.^276,295 The response to many of these concerns can be determined by the answer to the following question: Does the risk to an infant from a xenobiotic exposure via breast milk exceed the benefit of being breastfed?¹⁸³

Pharmacokinetic factors determine the amount of xenobiotic available for transfer from maternal plasma into breast milk; only free xenobiotic can traverse the mammary alveolar membrane. Most xenobiotics are transported by passive diffusion. A few xenobiotics, such as ethanol and lithium, are transported through aqueous-filled pores. The factors that determine how well a xenobiotic diffuses across the membrane are similar to those for other biologic membranes such as the placenta: MW, lipid solubility, and degree of ionization.

Large-molecular-weight xenobiotics, such as heparin and insulin, do not pass into breast milk. Lipid solubility is important not only for diffusion but also for xenobiotic accumulation in breast milk because breast milk is rich in fat, especially breast milk that is produced in the postcolostral period (~3–4 days postpartum). With a pH near 7.0, breast milk is slightly more acidic than plasma. Consequently, xenobiotics that are weak acids in plasma exist largely as ionized molecules and cannot be easily transported into milk. Conversely, xenobiotics that are weak bases exist in plasma largely as nonionized molecules and are available for transport into breast milk. In breast milk, ionization of the weak base xenobiotics occurs, and the xenobiotics are concentrated as a result of ion trapping. In other words, xenobiotics that are weak bases could concentrate in breast milk. Whereas sulfacetamide (pK_a of 5.4, a weak acid) has a concentration in plasma 10 times its concentration in breast milk, sulfanilamide (pK_a of 10.4, a weak base) is found in equal concentrations in both plasma and breast milk.¹⁸³

The net effect of these physiologic processes is expressed in the milk-to-plasma (M/P) ratio. Xenobiotics with higher M/P ratios have relatively greater concentrations in breast milk. The M/P ratio does not, however, reflect the absolute concentration of a xenobiotic in the breast milk, and a xenobiotic with a high M/P ratio is not necessarily found at a high concentration in the breast milk. For example, morphine has an M/P ratio of 2.46 (is concentrated in breast milk), but only 0.4% of a maternal dose is excreted into the breast milk.¹⁶ In general, for most xenobiotics, approximately 1% to 2% of the maternally administered dose is presented to the infant in breast milk.¹⁸³

The M/P ratio has several limitations. It does not account for differences in xenobiotic concentration that result from (1) repeat or chronic dosing, (2) breastfeeding at different times relative to maternal xenobiotic dosing, (3) differences in milk production during the day or even during a particular breastfeeding session, (4) the time postpartum (days, weeks, or months) when the measurement is made, and (5) maternal disease.

While being cognizant of the limitations, a spot breast milk xenobiotic concentration or a concentration estimate based on the M/P ratio allows a simplistic estimation of the quantity of xenobiotic to which an infant is exposed, assuming a constant breast milk concentration:

The effect of this dose on the infant depends on the bioavailability of the xenobiotic in breast milk, the pharmacokinetic parameters that determine xenobiotic concentrations in the infant, and the infant’s receptor sensitivity to the xenobiotic. These parameters are often different in neonates than in adults and could lead to xenobiotic accumulation; generally, absorption is greater, but metabolism and clearance are reduced.¹⁵ These effects are exaggerated in premature infants.^264,273 The amount of most xenobiotics delivered to infants in breast milk is usually adequately metabolized and eliminated.¹⁸³

Many of the above considerations are theoretical, and the number of specifically contraindicated xenobiotics is quite small.²⁸³ Published guidelines on the advisability of breastfeeding during periods of maternal therapy are generally based on the expected effects of full doses in the infant or on case reports of adverse occurrences. Interestingly, when the reports of adverse effects were reviewed, 37% of cases were in infants younger than 2 weeks, 63% were in infants younger than 1 month, and 78% were in infants younger than 2 months; 18% were in infants 2 to 6 months old; and only 4% were in infants older than 6 months.¹² It seems, therefore, that adverse effects are most likely to be observed in the first few weeks of life, when an infant’s metabolic capacity is significantly less than that of an older infant, child, or adult.^8,75

For most xenobiotics, a risk-to-benefit analysis must be made. For example, lithium is transferred in breast milk and leads to measurable, although subtherapeutic, serum concentrations in a breastfed infant. Although the effects of such exposure to lithium are unknown, many practitioners believe that the benefit of treating a mother’s bipolar illness outweighs the potential risk to the infant.^183,294 There is also a small increased risk of carcinogenicity associated with exposure to some environmental xenobiotics through breast milk.²⁷⁶

Similarly, the breastfed infant of a woman who smokes is exposed to nicotine and other tobacco constituents, both by inhalation and via breast milk. Although such a child is at increased risk for respiratory illness as a result of exposure to tobacco smoke, some of the risk may be mitigated by breastfeeding.^271,283

In most cases, women do not need to stop breastfeeding while using pharmaceuticals, such as most common antibiotics. However, “compatibility” with breastfeeding is generally based on a lack of reported adverse effects, which reflects limited clinical experience with a particular xenobiotic in breastfeeding patients. Therefore, in the setting of limited information, exposure to a xenobiotic through breast milk should be regarded as a small potential risk, and the infant should receive appropriate medical follow-up. Not all “compatible” medications are safe in all situations. For instance, phenobarbital can produce CNS depression in an infant if the mother’s serum concentration is in the high therapeutic or supratherapeutic range, which often occurs while dosage adjustments are being made. Such a concentration may or may not produce CNS depression in the mother. Nalidixic acid, nitrofurantoin, sulfapyridine, and sulfisoxazole, although generally safe, can cause hemolysis in a breastfed infant with glucose-6-phosphate dehydrogenase deficiency.

When women use substances postpartum, there is delivery of some xenobiotic to the infant via breast milk, and there are rare case reports of infants experiencing adverse effects.⁵⁵ Because of these possible direct effects on the baby, as well as the detrimental effects on the physical and emotional health of the mother and on the caregiving environment, the use of substances such as cocaine, methamphetamine, and heroin during the breastfeeding period is discouraged, and women actively using substances are discouraged from breastfeeding.²⁸³ However, women who are or have been abstinent from substance use and are participating in a treatment program are generally encouraged to breastfeed their infants.^4,183

Although ethanol is not specifically contraindicated for breastfeeding mothers, decreased milk production and adverse effects in infants are noted with maternal consumption of large amounts of ethanol.

Questions sometimes arise regarding possible lead exposure during breastfeeding. Lead crosses into breast milk from blood and is the most likely source of lead exposure for most breastfeeding infants.¹⁹⁶ Approximately 1% of US women between the ages of 15 and 49 years have blood lead concentrations greater than 5 mcg/dL.⁹² Some estimates suggest that breast milk concentrations are less than 3% of the maternal blood lead; therefore, if the maternal lead concentration is 5 mcg/dL, then the amount of lead delivered to the infant could be 1.5 mcg/L of breast milk.^124,170 This represents a relatively small exposure. In a sample of breastfed babies from one US city, the mean infant lead concentration was less than 3 mcg/dL, the highest concentration was 8 mcg/dL, and only 7.8% of values were greater than 5 mcg/dL.¹⁹⁶ The current consensus is that the benefits of breastfeeding outweigh the relatively small exposure to lead in breast milk.³⁴⁷ Blood lead screening for most women is not recommended.

There are, however, some subpopulations of pregnant women at increased risk of elevated environmental lead exposure for whom blood lead screening is recommended. Risk factors for significant lead exposure in pregnant women include recent immigration, pica practices, occupational exposure, poor nutritional status, culturally specific practices such as the use of traditional remedies or imported cosmetics, and the use of traditional lead-glazed pottery for cooking and storing as well as those involved with renovation or remodeling in older homes.⁹²

All women should have an assessment of risk for environmental lead exposure. Women at increased risk should have blood lead screening. When possible, pregnant and postpartum women with blood lead concentrations of 5 mcg/dL or higher should be removed from occupational or environmental lead sources and discouraged from practices or activities that result in increased exposure.⁹² In addition, some evidence suggests that dietary supplementation of calcium can reduce the mobilization of lead in postpartum women.¹³³ In cases of elevated maternal lead concentrations, decisions regarding breastfeeding should be made on an individual basis.

In 2007, after the death of a breastfeeding 13-day-old infant whose mother was using codeine, the US FDA issued a public health advisory regarding the use of codeine by breastfeeding women.³³³ In the initial case, the mother was found to be compound heterozygous for a CYP2D6*2A allele and a CYP2D6*2x2 gene duplication and therefore an “ultrametabolizer” of codeine.¹⁶⁶ In other words, codeine was metabolized to morphine at an exaggerated rate, and the infant ingested a high dose of morphine via the breast milk. The ultrametabolizer phenotype is present in up to 10% of whites and up to 30% of Ethiopians, North Africans, and Saudi Arabians.³³³ In addition, both the mother and the infant were homozygous for the UGT2B7*2 allele, which leads to elevated concentrations of morphine-6-glucuronide, an active metabolite.

Another decreased-function gene variant of ABCB1 has a role in this codeine toxicity. ABCB1 codes for a P-glycoprotein involved in transporting morphine out of the CNS. Decreased gene function would lead to increased accumulation of morphine in the CNS and potentially increased toxicity.³⁰⁶

Decisions on breastfeeding should be made with the informed involvement of the woman; her physicians; and when necessary, a consultant with special expertise in this field. Guidelines are available from several sources.^39,183

TOXICOLOGIC PROBLEMS IN NEONATES

Physiologic differences between adults and newborn infants affect xenobiotic absorption, distribution, and metabolism.^15,160,341 Appropriate administration of xenobiotics to newborn infants therefore requires an understanding of the appropriate developmental state for medication dosing and pharmacokinetics. Even so, approximately 8% of all medication doses administered in NICUs are up to 10 times greater or lesser than the dose ordered,⁶⁴ and as many as 30% of newborns in NICUs sustain adverse drug effects, some of which are life threatening or fatal.¹⁶ Pharmacokinetic differences between adults and newborns account for some cases of unanticipated xenobiotic toxicity that occur in newborn infants.

Gastrointestinal absorption of xenobiotics in neonates is generally slower than in adults.^15,160,341 This delay is related to decreased gastric acid secretion, decreased gastric emptying and transit time, and decreased pancreatic enzyme activity. The GI environment of newborns and young infants allows the growth of Clostridium botulinum and the subsequent development of infant botulism (Chap. 38). Infantile botulism is reported in infants several weeks of age.^148,326

Although it is uncommon, cutaneous absorption of xenobiotics is sometimes a route of toxic exposure in a newborn.^87,282 Aniline dyes used for marking diapers are absorbed, causing methemoglobinemia,²⁸² and contaminated diapers were responsible for one epidemic of mercury poisoning.²² The absorption of hexachlorophene antiseptic wash has led to neurotoxicity with marked vacuolization of myelin seen microscopically.^180,211,304 The dermal application of antiseptic ethanol has caused hemorrhagic necrosis of the skin of some premature infants. Iodine antiseptics have led to hypothyroidism in mature newborns.⁴⁹ An increased potential for absorption and toxicity has followed the application of corticosteroids^112,280 and boric acid⁸⁵ to the skin of children with cutaneous disorders.

Other routes of exposure have led to clinical poisoning. Several children inhaled talcum powder and died from acute respiratory distress syndrome.^40,229 Inhalation of mercury from incubator thermometers is a potential risk.¹¹ One child died after the ophthalmic instillation of cyclopentolate hydrochloride.²⁴

Because of differences in total body water and fat compared with adults, the distribution of absorbed xenobiotics differs in neonates.^15,160,341 Water represents 80% of body weight in a full-term baby compared with 60% in an adult. Approximately 20% of a term baby’s body weight is fat compared with only 3% in a premature baby. The increased volume of water means that the volume of distribution for some water-soluble xenobiotics, such as theophylline and phenobarbital, is increased.

Protein binding of xenobiotics is lower in newborns compared with adults: the serum concentration of proteins is lower, there are fewer receptor sites that become saturated at lower xenobiotic concentrations, and binding sites have decreased binding affinity.^15,160,341 Protein binding has potential relevance with respect to bilirubin, an endogenous metabolite that at very high concentrations causes kernicterus; bilirubin competes with exogenously administered xenobiotics for protein binding sites. In vitro, certain xenobiotics, such as sulfonamides and ceftriaxone, displace bilirubin from protein receptor sites, which might increase the risk of kernicterus, although this has not been clinically demonstrated. Conversely, bilirubin itself displaces other xenobiotics, such as phenobarbital or phenytoin, leading to increased plasma xenobiotic concentrations.

Newborn infants have decreased hepatic metabolic capacity compared with adults, which can lead to xenobiotic toxicity.^15,160 For example, caffeine, used in the treatment of neonatal apnea, has an extremely prolonged half-life in newborns because CYP1A2 has only 5% of the normal adult activity.¹⁵ Except for CYP1A2, most of the CYP isoenzymes reach approximately 25% of adult activity in newborns by about 1 month of age.

Two disease clusters related to immature metabolic function are described. The “gasping baby syndrome,” characterized by gasping respirations, metabolic acidosis, hypotension, CNS depression, convulsions, kidney failure, and occasionally death, is attributed to high concentrations of benzyl alcohol and benzoic acid in the plasma of affected infants.^10,43,113 Benzyl alcohol, a bacteriostatic, was added to IV flush solutions and accumulated in newborns after repetitive doses. The high concentrations of benzoic acid could not be further metabolized to hippuric acid by the immature liver. Immature glucuronidation in neonates is responsible for the “gray-baby syndrome” after high doses of chloramphenicol (Chaps. 31, 46 and 54).¹³⁸

The umbilical vessels are a common site of vascular access in sick neonates. Because blood drains into the portal vein, it is possible that IV medications experience a “first-pass” effect, although whether this route of xenobiotic administration affects metabolism or clearance has not been well studied. Most functions of the kidney, including glomerular filtration rate (GFR) and tubular secretion, are relatively immature at birth;¹⁵ the GFR of a newborn is approximately 30% of that of an adult. Xenobiotics such as aminoglycosides and digoxin are excreted unchanged by the kidney and therefore depend on glomerular filtration for clearance. Dosing of these xenobiotics in a newborn must account for these differences.

An interesting association has been made periodically over the years between the use of erythromycin, particularly in the first 2 weeks of life, and idiopathic hypertrophic pyloric stenosis.^{62,140,204,288,314} Although erythromycin is known to interact with motilin receptors in the antrum of the stomach, no specific etiology has been defined.¹³¹

Very little specific information is available to guide clinicians in the management of xenobiotic poisoning in newborn infants. In the case of cutaneous exposure to a xenobiotic, absorption is probably already complete by the time toxicity is noted. Nonetheless, the xenobiotic should be discontinued and the skin surface should be decontaminated. In the case of oral exposure, gastrointestinal decontamination is not generally performed in neonates, and neonates are at increased risk of fluid, electrolyte, and thermoregulatory problems after gastric lavage or the use of cathartics. Multiple-dose activated charcoal was used in a 1.4-kg, 2-week-old premature infant to treat iatrogenic theophylline toxicity.³⁰⁰ Hemodialysis, hemoperfusion, and exchange transfusion can be used in neonates to treat xenobiotic toxicity but require care in specialized units that experience with these extra-corporeal modalities (Chaps. 6 and 31).

SUMMARY

• Human embryos and fetuses are exposed to xenobiotics through the placenta of the pregnant woman; the neonates are exposed to xenobiotics via breast milk.

• Xenobiotic effects on developing humans include both congenital malformations as well as neurobehavioral abnormalities that manifest later in life.

• The use of xenobiotics in a pregnant or breastfeeding woman is a complex area of medical practice and presents clinicians with potentially difficult management decisions regarding the benefit of therapy to the mother and the risk to the mother or fetus of xenobiotic exposure.

• The goal is to optimize benefit for both the mother and the child but at least to minimize the risk to the fetus. In almost all cases, the primary approach is to fully and appropriately treat the mother.

• Appropriate management of many of the potential problems is facilitated by the coordinated efforts of obstetricians, perinatologists, neonatologists, pediatricians, and toxicologists.

REFERENCES

INTRODUCTION

Poisoning remains a significant but preventable cause of pediatric injury. Phone calls to poison control centers regarding pediatric exposures are more frequent than those regarding any other age group. Pediatricians have been leaders for 60 years in helping to establish and promote the study of medical toxicology, supporting the use of regional poison control centers and promoting the principles of poison prevention. Although the basic approaches to the medical management of toxicologic problems outlined in Chaps. 3 and 4 are generally applicable to both children and adults, some issues such as child abuse by poisoning are of particular concern. This chapter provides a perspective on the application of generally accepted toxicologic principles to children.

EPIDEMIOLOGY

To understand the magnitude of pediatric poisoning and its impact, epidemiologists examine multiple parameters, such as exposure, morbidity, mortality, and cost; however, these parameters are often difficult to measure accurately. An important source of information on the extent and effects of poisoning exposures in the United States is the American Association of Poison Control Centers’ (AAPCC’s) National Poison Data System. Every year, the AAPCC compiles standardized data collected from poison control centers throughout the United States; the 2015 annual review includes information submitted by 55 regional poison control centers. In the following discussion, comments on AAPCC data refer to cumulative information from the previous 5 published reports covering the years 2011 to 2015 (Chap. 130).

Each year, the AAPCC reports approximately 1.1 million potentially toxic exposures in young children from birth to 6 years of age that account for almost half (48%) of all reported exposures in the AAPCC database. Pediatric exposures involving all children and adolescents younger than 19 years account for 61% of all reported exposures. Of the reported pediatric exposures in children and adolescents, young children (younger than 6 years) account for 78%, school-age children (between 6 and 12 years) account for 10%, and adolescents (between 13 and 19 years of age) account for 12%. Females represent 43% of the reported poisoning exposures in young children and 58% of the reported exposures among adolescents.

Among the AAPCC-reported xenobiotic exposures in young children, 99% are coded as unintentional. There is some controversy regarding the use of the term unintentional with respect to childhood poisoning. A toddler quite purposefully intends to ingest a substance but does not intend to injure or harm him- or herself. Unsupervised is a term preferred by some epidemiologists to better describe the etiology of these exposures.¹¹⁰ In contrast, 39% of reported adolescent exposures are coded as unintentional, and 56% of adolescent exposures are coded as intentional. This shift in intentionality is due to an increased frequency of adolescent intentional substance use, abuse, and suicidal exposures. This high frequency of intentional poisoning in adolescents has been reported by others.^36,141,183

Table 31–1 shows the leading causes of AAPCC-reported exposures in children under 6 years. Exposures in this age group include xenobiotics that are commonly found around the house, such as cleaning products, cosmetics, plants, hydrocarbons, and insecticides as well as pharmaceuticals. It is important to understand that Table 31–1 lists the most commonly reported exposures, but that not all of these reported exposures are confirmed exposures and not even all confirmed exposures lead to serious morbidity or mortality. For example, children are most frequently exposed to cosmetic products; however, the amount ingested is typically small and most cosmetics manufactured in the United States are nontoxic, making the outcome of exposure minor.

Clinically significant exposures are unusual in children younger than 6 months. When they do occur, they most commonly result from the inadvertent administration of an incorrect drug or drug dose by a parent,^{23,62,110,126} intentional administration of a drug by a parent or sibling,^45,78 or passive exposure (eg, to the smoke of “crack” cocaine or phencyclidine).^{17,77,128,158,191} Any poisoning in a child younger than one year of age should be carefully evaluated for possible child abuse or neglect (see later discussion).⁷⁸

Several characteristics typically associated with ingestions by young children differentiate them from ingestions by adolescents or adults: (a) they are without suicidal intent; (b) there is usually only one xenobiotic involved; (c) the xenobiotics are usually nontoxic; (d) the amount is usually small; and (e) toddlers usually present for evaluation relatively soon after the ingestion is discovered, generally within 1 to 2 hours. The peak age for childhood poisoning is between 1 and 3 years.^{23,31,110,155} Unintentional ingestion is unusual after age 5 years although when it occurs it is sometimes the result of mistakenly consuming a xenobiotic from a mislabeled container.²⁶ Some intentional ingestions between the ages of 5 and 9 years result from intrafamilial stress or suicidal intent. After age 9 years and through adolescence, overdose or poisoning frequently results from either suicidal gestures or suicidal attempts or from the adverse effects of alcohol or substance use. This previously discussed age-dependent shift in intentionality toward substance use and abuse and suicide is associated with a significant change in outcome. Only 1% of exposed children under 6 years have a recorded outcome as moderate, major, or death compared to 17% of adolescents (Table 31–2).

The peak age for hospitalizing young children exposed to xenobiotics is between 1 and 3 years, reflecting the peak age of exposure. Whereas hospitalization of children younger than age 2 years typically results from exposure to nonpharmaceutical xenobiotics, children older than 2 years as well as adolescents are more commonly exposed to pharmaceuticals.^55,63,183 Adolescents are also thought to be more frequently hospitalized than children after exposure to xenobiotics but this is a subject of debate.⁶³ This is a result of the need for medical management and, often, psychiatric hospitalization. As many as 30% of children who experience one ingestion will experience a repeat ingestion; adolescents are particularly prone to recidivism.^63,86 Children who ingest poisons are also at a greater risk for other types of injuries.^15,53

Another source of data related to poisoning morbidity is the National Electronic Injury Surveillance System (NEISS) online database maintained by the US Centers for Disease Control and Prevention (CDC), which provides information on emergency department (ED) visits and hospitalizations.³² For the period 2005 to 2014, the CDC estimates approximately 45,000 ED visits and 3,400 hospitalizations (8%) per year for young children 0 to 5 years who are exposed to a xenobiotic. Children 6 to 12 years old account for approximately 10,000 ED visits and 1,000 hospitalizations (10%) per year and adolescents 13 to 19 years old account for approximately 109,000 ED visits and 21,000 hospitalizations (19%) per year.

In a published review of NEISS data related to medication exposures in young children, approximately one-half of the ED visits were the result of unsupervised medication exposures and one-half were the result of adverse drug events after caregiver medication administration.¹¹⁰ Approximately one-half of the medications were prescription pharmaceuticals and the other half were over-the-counter products. The most commonly involved prescription medication categories were opioids, benzodiazepines, antidepressants, beta adrenergic antagonists, and amphetamines. Ninety-two percent of all over-the-counter exposures involved acetaminophen (APAP), cough and cold products, ibuprofen, and diphenhydramine. The admission rate in this population was 18.5% for unsupervised ingestions and 6.0% for caregiver-administered medication events.

Although the AAPCC reports overall outcome related to age, it does not specifically stratify outcome of exposure to individual xenobiotics by age. In a multiyear review published in 1992, the AAPCC did report the xenobiotics that caused the greatest number of major and fatal effects in young children. The categories associated with the greatest number of major effects included cleaning substances, cardiovascular drugs, hydrocarbons, sedative/hypnotics, and antidepressants.¹⁰⁷ A more recent study reported a shift in the substances with highest morbidity requiring pediatric intensive care unit (PICU) admissions. PICU admissions following unintentional ingestions were more frequently due to buprenorphine/naloxone, cardiac medications, and alpha-2-agonists. However, PICU admissions following intentional ingestions were due to medications from 43 different drug classes; patients frequently took their own medications, over-the-counter medications, or substances they obtained from a friend often with suicidal intent.⁵⁴

Other national and international reports of hospitalized patients include a similar distribution of xenobiotic exposures that cause significant morbidity and hospitalizations.^4,10,33,197 However, in rural areas of many developing countries, kerosene and pesticides are the leading causes of xenobiotic-related hospitalizations, and the spectrum of pharmaceutical exposures is often different (Chap. 136).^{2,72,74,127,134,136}

Poisoning accounts for approximately 2% of young childhood and 7% of adolescent injury-related deaths in the United States.³² Based on information from death certificates filed in state vital statistics offices as well as demographic information provided by funeral directors, the CDC reports 1023 poisoning deaths in children younger than 6 years of age from 2005 to 2014 for an average of about 102 per year. These deaths represent a 78% decrease from the 456 deaths reported by the CDC in 1959.³¹ There are several factors responsible for this decrease, including improved poisoning prevention strategies such as child-resistant closures and improved medical care. In addition, some industrial or pharmaceutical products previously found around the home have been replaced with less toxic products or were reformulated to safer concentrations.²⁸ It is also possible that there has been a decrease in reporting (Chap. 130).

Although the AAPCC data provide a remarkable amount of epidemiologic information, there are questions about the accuracy of the data.^75,76,190 For example, as mentioned earlier, the CDC reported 1023 poisoning-related deaths in children younger than 6 years of age from 2005 to 2014. During the same period, the AAPCC reported 253 deaths in young children. A primary weakness of the AAPCC data is due to its voluntary reporting method of ascertainment and it must be recognized that many significant poisonings go unreported to poison control centers. In Rhode Island, only 45 of 369 poisoning deaths were reported to the regional poison control center¹⁰⁵ (Chap. 130).

According to the 2013 AAPCC National Poison Database System pediatric fatality review, 47% of the reported pediatric fatalities resulted from unintentional exposures. Of these 47% were environmental exposures, mostly carbon monoxide poisoning, 40% were xenobiotic ingestions, and 10% were therapeutic errors.¹¹¹ Fifty-three percent of AAPCC-reported pediatric fatalities were the result of intentional exposures primarily related to substance use, abuse, and suicidal intent.

There are some significant etiologic differences between children and adolescents, particularly with regard to the lethality of xenobiotics. According to the 2013 NPDS pediatric fatality review, the xenobiotics causing the greatest number of childhood deaths were carbon monoxide, household products (gasoline, batteries, cleaners, laundry detergent, lamp oils), and pharmaceuticals, with opiates (buprenorphine and methadone) accounting for 10% of the unintentional childhood deaths. In contrast, the xenobiotics causing the greatest number of adolescent deaths were “designer drugs” (psychogenic phenylethylamines and synthetic cannabinoids), prescription opioids (methadone, oxycodone, hydrocodone, buprenorphine), and antidepressants. Unfortunately, there is an increasing trend in exposures of adolescents to these substances, and although the overall number of actual deaths from use, abuse, and suicide remain too small to comment on statistically, they are of concern.

The most notable difference between the xenobiotics causing pediatric morbidity and mortality today and studies from the 1960s and 1970s is that salicylates are no longer a leading cause of reported poisoning morbidity and mortality.^40,44 This change is likely related to a combination of factors, including federal regulations requiring child-resistant closures, the 81-mg dose form replacing higher doses as the most common formulation of aspirin prescribed, as well as the overall decreased use of aspirin in children (Chap. 37).^{18,38,84,129,147}

Poisoning also has an economic cost. An ED visit for a poisoned child costs $1,077, on average, while the mean charge for hospitalization is $11,792, which can vary significantly based on the length of stay and outcome. Based on the estimated number of poisoned children seen in the ED each year, the total US health care cost for emergency care would be estimated as $162.3 million dollars.¹³¹ According to the CDC in 2010, for fatal poisonings of children and adolescents, the average medical cost was approximately $6,000, but the average work loss cost was $1.9 million for a total cost of $1.9 billion.³²

BEHAVIORAL, ENVIRONMENTAL, AND PHYSICAL ISSUES

An oversimplification of the etiology of childhood poisonings would be the formulation that unsupervised toddlers exploring their environment inadvertently ingest xenobiotics. However, this approach ignores the complex interplay of factors that contribute to most ingestions in children.

One approach to understanding injury causation uses the Haddon Matrix.⁷³ Applying this model to poisoning, there are 3 interacting factors: the patient (host), the xenobiotic (agent), and the environment. These factors interact during 3 phases: preinjury, injury, and postinjury. The factors themselves contribute to the likelihood, nature, magnitude of, and host response to an injury. Modification of these factors helps reduce the frequency or severity of an injury.

For example, if a 2-year-old child finds a tablet there is a risk of ingestion. Storing the tablets in a child-resistant container (CRC) or storing the container out of reach can help to reduce the risk of ingestion. In this case, both the agent and the environment are modified in the preinjury phase. However, both of these strategies are considered active prevention strategies, as they require the caregiver to actively replace the top back on the container or actively place the container back on a high shelf. Another method of preinjury prevention involves passive prevention. For example, if the tablet is an iron-containing multivitamin, limiting the amount of iron in each tablet would make the formulation of the product safer overall. In this regard, the availability of a safer product is passive for the user but active for the manufacturer or for regulators.²⁹

The injury phase covers both the ingestion and the initial pathophysiologic host response. Again, particular factors determine the nature and extent of injury. Continuing the example above, if the tablet is a 325-mg APAP tablet, the ingestion will not lead to injury. However, if the tablet is a 0.2-mg clonidine tablet, there is a higher likelihood of toxicity.

The postinjury phase is concerned with both the ongoing host response and the medical management of the patient who is poisoned. In this phase, it would be determined whether the 2-year-old child with a clonidine ingestion is manifesting signs of toxicity, such as coma or hypotension; whether the child requires treatment with activated charcoal (AC), intravenous (IV) fluid, or naloxone; and whether the child requires hospital admission.

This paradigm is only a model. In reality, it is often difficult to examine any individual factor independently, and the relative contributions of these factors are not well defined. Nonetheless, consideration of the individual factors of host, agent, and environment allows us to focus on several relevant aspects of poisoning in children.

Childhood and adolescence are times of tremendous growth and development.²⁰⁵ Some of these physical and social changes place children and teenagers at increased risk for poisonings. By 7 months of age, an infant can sit up and can pivot to grab an object; by 9 to 10 months of age, most infants can creep and crawl; by 15 months of age, most toddlers are walking quite competently and eagerly exploring. Between 9 and 12 months of age, a skillful pincer grasp with the thumb and forefinger develops that allows the child to pick up small objects. Throughout this period, one of the child’s primary sensory experiences is sucking on or gumming objects that are placed in the mouth. Thus, the combination of 3 developmental skills—the ability to move around the home and depart the immediate view of a guardian, the ability to pick up and manipulate small objects, and the tendency of children to put things in their mouths—places them at risk for poisoning.

As children develop socially, they desire to become more like their parents and they tend to imitate behaviors such as taking medicine or using mouthwash. Children are typically taught that medicine is good for them when they are sick. Many children’s medicines are sweetened and flavored to make them more palatable, and many parents inappropriately encourage their children to take medicines by telling them “it tastes like candy.” Children are observed “making tea” from plants or “making pizza” with mushrooms from the yard.²⁶

As children become more mobile, agile, and curious, xenobiotics that were previously outside their reach become accessible even when stored in difficult-to-reach places. Evidence suggests that some parents underestimate the developmental skills of their children.⁵³ Although 87% of parents believe unintentional injuries and poisonings are preventable, only 68% actually take action such as using safety belts or car seats. These parents also consider “being careful” as an effective preventive action and only 10% take any action toward removing or securing poisons.⁵² Unfortunately, a primary motivator to shift parental perception of risk and precipitate action occurs only after a child poisoning event or a near miss when it is too late. Additionally, there is often the risk of perceived safety if an event is not associated with any morbidity.^66,130

As previously discussed, some of the reasons why a child ingests a pill are that it is there, it looks and maybe tastes like candy or food, or the child is mimicking the behavior of a parent who ingests medicines or vitamins to cure illness and improve health. However, these reasons may not be sufficient to explain why xenobiotic exposures occur. Another aspect of poisoning that must be considered is the interaction between the child’s temperament and the social environment.

Many authors have tried to identify psychosocial predictors for childhood poisoning in general and for repeat poisoning in particular.^{21,56,87,90,173} As many as 30% of children repeatedly ingest xenobiotics, frequently the same xenobiotic. Factors associated with single and repeat episodes of childhood poisoning include hyperactivity, impulsive risk-taking behavior, rebelliousness, and negativistic attitude. Other factors seem to be associated more with the quality of supervision by parents or guardians, who themselves are experiencing medical illnesses, depression, or social isolation.⁹⁰ Finally, a stressful environment or a major social problem also contributes.^167,174

With regard to the xenobiotic, a number of issues affect the preinjury and injury phases, and modification of any one of the interacting factors of host, xenobiotic, or environment potentially prevents or reduces the severity of injury. By decreasing the inherent toxicity of household products available around the house, the likelihood of injury is reduced after exposure to one of these products. For example, less toxic rodenticides such as warfarin replaced more toxic ones such as thallium or sodium monofluoroacetate and relatively nontoxic paradichlorobenzene mothballs have largely replaced the relatively more toxic camphor-containing mothballs.

It is also possible to reduce the likelihood of ingestion by making a xenobiotic unpalatable. Denatonium benzoate is an aversive bitter xenobiotic that is added to some liquids such as windshield washer fluid and antifreeze to prevent unintentional poisoning.⁸⁵ However, some trials show that whereas older children respond negatively to these xenobiotics with the first taste, younger children have already ingested 1 to 2 teaspoonfuls before being deterred by the bitter flavor.^20,166 This is an important consideration because even a small amount of a xenobiotic such as methanol can be toxic (see later discussion). The actual efficacy of denatonium benzoate in poison prevention is largely unstudied.¹⁴⁸

The problem of unintentional ingestions is compounded by poison “look-alikes,” xenobiotics that resemble candy or food products.⁵⁷ Some common examples are ferrous sulfate tablets and vitamins that look like common candies and fuel oils that come in cans resembling soft drink containers. Many shampoos and dishwashing detergents are given lemon or strawberry scents and have pictures of fruits on the labels. More recently changes in marijuana legalization across the country have led to more edible marijuana products coming to the market. Children easily confuse marijuana cookies and candies with their traditional counterparts. In states where these products are legally sold, the number of poisoned children is climbing.^19,188 Children are not always able to distinguish poison “look-alikes” from real candies, fruits, and sodas, and they may be attracted to bright colors, pleasant smells, and appealing packages. Eliminating these “look-alikes” might prevent some unintentional ingestions.

Probably the most significant changes to xenobiotics themselves have occurred in the physical characteristics with regard to packaging and dispensing of pharmaceuticals and some other xenobiotics with child-resistant closures mandated by the Poison Prevention Packaging Act of 1970 (Chap. 1).^29,39 According to the act, “child-resistant” means that the packaging must be effective at preventing 85% of children from opening the package and yet allow 90% of adults to open the package. This legislation is credited with causing a significant reduction in morbidity and mortality caused by poisoning from aspirin and other regulated products, although this analysis was challenged.^38,146,186 Child-resistant closures are also credited with reducing the number of toxic exposures to kerosene.¹⁰¹

Although child-resistant containers are a significant deterrent to unintentional ingestions in toddlers, they are not completely effective; by definition, they allow 15% of children to gain access. Problems include the dispensing of pharmaceuticals in nonresistant containers, not properly closing child-resistant containers, leaving pharmaceuticals out of the child-resistant container, and transferring products to a different nonresistant container.^30,172 Up to 70% of potentially toxic pharmaceuticals are stored in non–child-resistant or in improperly functioning child-resistant containers. Several studies identified poor functioning of the closures when there is sticky liquid or pill residue around the top or in the screw threads of the child-resistant container.^86,94,195 A false sense of security associated with these closures leads some parents to be less compulsive regarding safe storage of the containers.^66,130 A double barrier, such as a unit-dose dispenser within a child-resistant container or a blister pack, is recommended for a few pharmaceuticals associated with significant poisonings such as iron and antidepressants.⁹⁴ Recently, flow restrictor valves were voluntarily added to some over-the-counter analgesic liquid products to help decrease the volume of liquid dispensed when the child-resistant closure is not fully secured or the child is able to breach the closure mechanism.¹⁰⁹

In 1997, the Food and Drug Administration (FDA) issued a regulation to package products with 30 mg or more of elemental iron per tablet in unit-dose packages such as blister packs.⁵⁹ The intent of this regulation was to reduce the likelihood of iron poisoning in children. Even before this regulation was instituted, the number of fatal childhood iron ingestions declined significantly; therefore, it is not known how much if any of the decrease was related to the mandated packaging changes. In any case, the rule was overturned in 2003, when it was determined that the FDA did not have the statutory authority to regulate a drug for the purpose of poison prevention.⁶⁰ As yet there is no evidence of a resurgence of iron-related fatalities.

A discussion of containers and storage naturally leads to a consideration of the third factor in the injury-causation model discussed earlier, the environment, which is particularly important in the preinjury and the injury phases. Approximately 80% of childhood pharmaceutical ingestions occur at home; most of the remainder occurs at the homes of grandparents, other relatives, and friends. At home, the medicine usually belongs either to the child or to a parent, although a significant number of medications, both at home and away from home, belong to a grandparent.^86,106 Grandparents, other relatives, and family friends without children at home often do not obtain or retain medications in child-resistant containers and are not as attentive to safe storage practices. Poison prevention education directed to these groups is particularly helpful.¹¹⁹

Medications are frequently kept in the kitchen or bedroom while they are being used.^86,195 In the kitchen, medications are stored in the refrigerator, on the table, or on the counter; in the bedroom, medications are left on a dresser or bedside table. A parent’s or grandparent’s belongings are another location where medications are commonly found. Interestingly, there are no significant differences in the storage practices in the homes of children who ingest and those who do not ingest medicines, so storage practices alone cannot predict the likelihood of childhood poisoning.^173,195

One important caveat relates to the storage of nonpharmaceutical xenobiotics, particularly those in liquid form. These types of xenobiotics should never be transferred for storage to familiar household containers, such as food jars or wine or soda bottles; stored in areas low to the ground such as beneath sinks; or kept in cabinets that do not have child-resistant locks. Both children and adults have been unintentionally exposed to xenobiotics, such as sodium hydroxide, pesticides, hydrocarbons, and potassium cyanide, that were stored in bottles in the refrigerator.¹⁸² Many of the kerosene exposures reported from developing countries occur because the kerosene is stored in water bottles, jugs, or other containers in easily accessible locations. When the weather is hot, children mistake the clear liquid kerosene for water.^1,2,165

HISTORY OF THE INGESTION

The appropriate management of any poisoned patient is influenced by the history of the exposure. Except in rare cases of child abuse, parents or guardians generally provide information to the fullest extent possible. As a rule, in the case of children, the xenobiotic and time of ingestion are known. However, the reported number of pills or volume of liquid ingested is rarely accurate.^80,150 Clues to the amount ingested are the number of pills or volume of liquid in a bottle before and after an ingestion, the number of pills set out on the night table, or the area of a spot of liquid after a spill. When symptoms are suggestive of poisoning but the history is inadequate, information about possible exposure outside of the home, such as with a babysitter, grandparent, friend, or other relative, should be obtained because approximately 15% of childhood poisonings occur outside the home.^86,138

In contrast, many adolescents are not forthcoming when relating the history of an intentional ingestion, especially when they are depressed, suicidal, or concerned about the response of their guardian or parent. When caring for these patients, the clinician must use the history provided but should remain skeptical about the reported type and number of xenobiotics ingested, as well as the time of ingestion.

When a child is suspected of being the victim of abuse or intentionally poisoned by a parent or guardian, the health care provider must ensure that (a) the history of the poisoning remains consistent over time and among people providing the details of the event, (b) the child’s clinical presentation is consistent with the history of the poisoning, and (c) the reported actions are consistent with the child’s developmental level.

GASTROINTESTINAL DECONTAMINATION

Chapter 5 is devoted to a complete discussion of gastrointestinal (GI) decontamination. This section reiterates and emphasizes only a few important points.

As described earlier, children generally ingest small quantities of a single xenobiotic. For most of these ingestions, gastric emptying is unnecessary. Some examples of nontoxic ingestions are eating a crayon or the leaf of a jade plant, licking the cap of a household bleach container, or swallowing 2 adult-strength APAP tablets.

Orogastric lavage is a method of gastric emptying occasionally indicated for the most serious ingestions. Orogastric lavage typically requires a very large bore 40-French tube in order to guarantee that the interior diameter and side-port holes are large enough to allow pill fragments to effectively pass back out through the tube. Small children, however, cannot generally tolerate such a large orogastric lavage tube, and a smaller tube is unlikely to be effective for removing large pills or fragments from the stomach of a small child. Additionally, placement of an orogastric tube is an unpleasant and frightening procedure for an infant or child. There is a risk of local trauma from tube placement, and there is the potential for more serious injury such as esophageal perforation. Also, many children vomit during placement of an orogastric tube, increasing the risk for aspiration. Therefore, the use of orogastric lavage is practically limited to adolescents with large potentially fatal ingestions. Orogastric lavage should never be used as a form of punishment or as a form of education. The patient should be intubated to protect the airway before orogastric lavage if the patient has a diminished gag reflex or a depressed level of consciousness.

Historically, administration of syrup of ipecac to poisoned patients was considered a primary emergency intervention and the availability of syrup of ipecac in the home was a primary tenet of pediatric anticipatory guidance. Since 2004 the American Academy of Pediatrics, the American Academy of Clinical Toxicology, and the European Association of Poison Control Centers and Clinical Toxicologists have all recommended against the use because of its unproven benefit, its adverse effects, and its abuse potential.^6,8

Activated charcoal (AC) is a current mainstay of poison treatment in EDs.^5,37 Children will generally drink AC when coaxed to do so if the AC is disguised in a baby bottle or soft drink container or sweetened with juice or sorbitol.¹⁹² A nasogastric or orogastric tube should rarely ever be inserted to administer AC in an awake patient. Placement of the tube, the presence of AC in the stomach, the effects of the xenobiotic, all may make the child vomit, making aspiration of AC or stomach contents a risk. For AC to be used safely in a patient who is comatose and who does not have a gag reflex, the patient should be intubated and the airway protected first, prior to insertion of an orogastric tube. Because of these risks, AC alone is unnecessary for a nontoxic or minimally toxic ingestion.

METHODS OF ENHANCED ELIMINATION

For consequential poisoning with xenobiotics such as methanol, ethylene glycol, salicylates, lithium, and theophylline, hemodialysis is the optimal technique to enhance elimination. It is feasible to perform these extracorporeal techniques on newborns or small infants in appropriately equipped centers with dedicated personnel. The primary limiting factor is the ability to obtain vascular access.^124,177,178 However, even large centers that routinely do hemodialysis in children will periodically not be able to manage very small infants. Peritoneal dialysis was used for the treatment of ethanol intoxication in a child,⁷¹ but this technique is not recommended.

Exchange transfusion is occasionally used to enhance xenobiotic elimination. This technique is potentially useful for a xenobiotic with a small volume of distribution when multiple-dose AC cannot be administered, the xenobiotic is poorly adsorbed to AC, and when access to specialized hemodialysis or hemoperfusion is not readily available. Exchange transfusion was used successfully for poisoning by salicylates^50,117 and theophylline.^16,135,161 Another xenobiotic for which exchange transfusion is a potential therapeutic alternative is chloral hydrate.¹¹

XENOBIOTICS THAT ARE TOXIC IN SMALL QUANTITIES

When children ingest even small quantities of toxic xenobiotics, they are potentially ingesting relatively large doses because of their small size. There are a number of xenobiotics that cause significant toxicity or rarely death with as little as one pill or one teaspoonful.^14,104 Table 31–3 lists these xenobiotics.

XENOBIOTICS THAT CAUSE DELAYED TOXICITY IN CHILDREN

Several xenobiotics warrant particular concern because their effects are frequently significantly delayed. Classic examples are atropine–diphenoxylate (Lomotil)^22,42,118 and sulfonylurea antidiabetics such as glipizide.^69,139,180 Both of these xenobiotics have the potential to cause serious morbidity with initial symptoms or recurrence of symptoms delayed by as many as 24 hours after ingestion.

Children with real or possible ingestions of Lomotil or a sulfonylurea should be admitted for observation and monitoring even if they are initially asymptomatic (Chaps. 36 and 47). With the advent of new modified-release formulations of calcium channel blockers and β-adrenergic antagonists, concern for delayed toxicity and possibly death has become even greater.¹²⁶

XENOBIOTICS THAT CAUSE UNUSUAL OR IDIOSYNCRATIC REACTIONS IN NEONATES OR CHILDREN

Benzyl Alcohol: Gasping Syndrome

Benzyl alcohol is a preservative added to some liquid pharmaceutical preparations (Chap. 46). IV flush solutions containing benzyl alcohol were implicated as the cause of the “gasping syndrome” in sick newborns; the syndrome includes severe metabolic acidosis, encephalopathy, respiratory depression, and gasping.⁷ The association was made when elevated concentrations of benzoic acid and hippuric acid—metabolites of benzyl alcohol—were found in infants with this syndrome.^27,65

Chloramphenicol: Gray Baby Syndrome

Chloramphenicol is a broad-spectrum antibiotic that was used in children because of its activity against Haemophilus influenzae. It was largely replaced by other antibiotics in the United States because of its association with aplastic anemia. When administered at high doses, chloramphenicol is associated with the “gray baby syndrome,” which includes abdominal distension, vomiting, metabolic acidosis, progressive pallid cyanosis, irregular respirations, hypothermia, hypotension, and vasomotor collapse. Although these effects occur primarily in premature newborn infants, they also occur in older children and adults (Chap. 54).

Gray baby syndrome is associated with serum chloramphenicol concentrations greater than 100 mg/L. Increased chloramphenicol concentrations result from (a) inadequate conjugation of chloramphenicol with glucuronic acid because of inadequate activity of glucuronyl transferase in the newborn liver and (b) decreased renal elimination of unconjugated chloramphenicol. The exact mechanism of toxicity is unknown; there is speculation that free radicals produced during the metabolism of chloramphenicol may interfere with mitochondrial function.⁸²

Ethanol: Hypoglycemia

Ethanol is a major constituent of many liquid preparations, such as mouthwash, vanilla flavoring, perfume, and hand sanitizers. Besides its well-known sedative–hypnotic effects, ethanol poisoning in children is associated with hypoglycemia. Typical ethanol metabolism to acetic acid results in a shift to higher NADH:NAD⁺ ratio. This redox state favors the conversion of pyruvate to lactate, diverting pyruvate away from the Krebs cycle and gluconeogenesis. This decreased source of energy and glucose is poorly tolerated in children, especially because of reduced hepatic glycogen stores. Ethanol-induced hypoglycemia increases the risk for seizure and exacerbation of other CNS effects induced by ethanol poisoning. There does not seem to be a definite blood alcohol concentration threshold for the development of hypoglycemia, which has been reported with blood alcohol concentrations as low as 20 mg/dL⁴¹ (Chap. 76).

Imidazolines and Clonidine: Central Nervous System Effects

Imidazolines such as tetrahydrozoline, oxymetazoline, xylometazoline, and naphazoline are nonprescription sympathomimetics used as nasal decongestants and conjunctival vasoconstrictors (Chap. 49). Clonidine is an imidazoline derivative used as an antihypertensive (Chap. 61). In small children, these xenobiotics cause severe central nervous system (CNS) depression, respiratory depression, bradycardia, miosis, and hypotension.^13,116,193 The presumed mechanism of action is by stimulation of central α₂-adrenergic and imidazole receptors. Although naloxone has been reported to reverse some of the CNS effects of clonidine, there are no reports of its successful use with the other imidazolines (Chap. 61).

HISTORICAL POISONINGS

There have been numerous times throughout history that children have been disproportionately affected by a toxicologic mass casualty incident.

Aniline Dye

In 1886, Rayner first described a series of newborn children on his maternity ward appearing dusky in color and somewhat limp. Six months later when a second series of newborns presented, he linked the symptoms to a delivery of diapers that were freshly stamped with aniline dye.⁸⁹ The dye was absorbed through the newborn skin causing symptoms of what is now known as methemoglobinemia. Despite reports and warnings, aniline continued to be used as a diaper marking dye for decades.^83,89,137 Today, we continue to see pediatric aniline-dye-induced methemoglobinemia from contaminated Indian Holi festival dyes and shoe dyes as well.^132,170

Sulfanilamide Elixir

In the late 1930s, the antibiotic medication sulfanilamide was used to treat infections such as streptococcal pharyngitis. Because of a demand for a liquid preparation, an elixir was rushed to market without adequate testing. Sulfanilamide itself is poorly soluble and many of the typical diluents were inadequate; however diethylene glycol functioned well as a solvent and was sweet tasting. Unfortunately diethylene glycol is severely toxic and causes acute kidney injury.⁶⁴ The use of diethylene glycol as a diluent for sulfanilamide caused more than 100 deaths including many children.^12,187 As a result and to avoid similar tragedies, the Food, Drug, and Cosmetic Act was passed in 1938 that requires animal testing of new drugs and data submission to the FDA before marketing to the public.

Unfortunately diethylene glycol continues to be used sporadically around the world with the same tragic effects. In 1996, in Haiti, there were 85 deaths of 109 children treated for acute kidney injury associated with the use of diethylene glycol–contaminated APAP syrup.¹³³ Again in 1998, in India, 33 out of 36 children who presented with acute kidney injury died after ingesting diethylene glycol–containing cough expectorant.¹⁶⁹ Whether use of diethylene glycol in these cases despite available information regarding its toxicity was out of ignorance or related to unethical financial corruption has not been resolved.

Melamine

Melamine has widespread uses, most commonly in the plastic dinnerware industry. In the trace quantities that may leach from dinnerware into food to be eaten, melamine is generally considered not toxic. However, in 2008, in China, an outbreak of infantile urolithiasis was associated with melamine-containing powdered infant formula.

The protein content of food is measured based on its nitrogen content. Melamine is more than 60% nitrogen by weight, and by adding melamine to a dry powdered formula the apparent protein content is falsely elevated. Unfortunately, some Chinese formulas were measuring more than 6,000 ppm melamine compared to less than 1 ppm in a comparable US product.¹⁷¹ More than 294,000 Chinese children were affected, leading to 51,900 hospitalizations and 6 deaths.

MEDICATION ERRORS

Ever since the publication of the Institute of Medicine’s report titled To Err Is Human in 1999, increasing attention is being paid to the issue of medical errors in medicine.⁹⁵ Although most of the research regarding medication errors has focused on adults (Chap. 134), this problem also affects children. Remarks in this section are generally limited to the pediatric literature.

Approximately 276,600 exposures each year are classified by the AAPCC as therapeutic errors. For 2011 to 2015, there were a total of 7 deaths attributed to therapeutic errors. Of children younger than the age of 6 years, approximately 40% of the errors resulting in severe injury or death occurred in children younger than the age of 1 year.¹⁸⁴

Medication errors occur at any phase of a process that includes ordering, order transcription, pharmacy dispensing, preparation and administration of the medication, and monitoring of medication effects. In fact, the same types of errors occur at multiple points in the process. Table 31–4 provides some examples of errors.

Most of the analyses of medication errors have occurred in inpatient settings. The reported frequencies of medication errors vary widely—from 0.47% to 5.7% of written orders and from 0.51 to 157 per 1,000 patient-days.^{58,81,88,91,153}

The variance largely depends on whether the definition of “error” does or does not include prescribing errors, regardless of whether or not they are corrected, and whether potential, or only actual, adverse drug events are included. The reported frequencies also vary depending on whether there is active case finding or whether there is only voluntary reporting.

In a 2001 study of pediatric inpatients in which orders were actively monitored for a 6-week period, 5.7% of 10,778 prescriptions had errors in the order for the medication, transcription of the order, dispensing or administration of the medication, or monitoring of medication effects (56 per 100 admissions, 157 per 1,000 patient-days).⁹¹ Overall, 1.1% of these errors could have potentially caused an adverse effect (10 per 100 admissions, 29 per 1,000 patient-days). Eighty-four percent of the errors occurred during the ordering or transcription phase, so most of the errors were intercepted and corrected before drug administration. There were 26 true adverse drug events, but only 5 were considered preventable errors (0.05%, 0.52 per 100 admissions, 1.8 per 1,000 patient-days). Although the overall error rate was similar to that reported by the same group in a study of adults, the rate of errors that could potentially have caused harm was 3 times greater; 41% of the potentially harmful errors were not intercepted. However, one review suggests that the prescription error rate is higher in adults than in children.¹⁰³

Generally, error rates are higher in intensive care units, where the sickest patients are cared for; such patients often receive multiple medications with complex administration regimens.^{58,81,91,140,153,185,194} Results similar to those from inpatient settings are reported in pediatric EDs.^99,159

Children are at increased risk of a medication error for several reasons: (a) someone other than the child administers the medication, so there is little opportunity to prevent or limit drug administration; (b) a young child cannot warn practitioners about possible problems such as allergies; (c) a young child cannot inform practitioners when he or she is experiencing an adverse event; (d) medication ordering and administration in children frequently requires dose calculations; (e) inexperienced practitioners are uncomfortable with pediatric dosing or related calculations; and (f) incorrect measurement or dilution of concentrated stock solutions may yield a small volume, which is not perceived as containing a relatively large dose of medication.^34,46,67

One of the most common errors is prescription, preparation, or administration of an incorrect dose, particularly in children, for whom almost every prescription requires knowledge of the patient’s weight and a calculation of a weight-based dose.^91,100,102 In addition, even the milligram per kilogram dose varies depending on the age of the patient or the diagnosis. Although pediatric doses are generally determined on a milligram per kilogram basis, if the weight is recorded in pounds and this weight is used in the calculation, there will be a built-in 2-fold error. Calculation errors also occur when drug preparation requires dilution of a concentrated stock solution or special compounding.¹⁸⁴ Further confusion arises when mg is written as or misinterpreted as mcg or μg in a calculation or vice versa.⁴³

When an extra zero is added or a required zero is omitted from calculations, written or verbal prescriptions, or in dispensed and administered medications, a 10-fold error occurs. These large errors are common and result in significant under- or overdosing; 10-fold errors are reported in testing scenarios, case series, and case reports.^{46,49,96,97,100,140,144,149,153} These errors are of particular concern because the risk of toxicity generally increases with significant overdose.

Because the causes of medication error are numerous and complex, the solutions must be multifaceted and interdisciplinary. The approach to the problem is contained within the field of human factors research and potentially requires changes in individual factors such as knowledge; environmental factors such as interruptions; and system problems such as how medications are ordered, stocked, and dispensed (Chap. 134).^95,179,189

Pediatric critical care areas, including the ED, benefit from having precalculated weight-based dosing charts available for resuscitation medications and for other commonly prescribed medications; this is a recommendation of the American Heart Association.⁹ Many clinical units have developed their own dosing schemes. Commercial products, such as the Broselow–Luten system, are also available and reduce the number of medication errors in simulated^{3,112,113,160} and actual resuscitations.¹⁵² There is some controversy related to the ability of these length-based commercial products to accurately predict the weights of American children because of the growing obesity epidemic. However, most resuscitation medications should be dosed on ideal body weight as estimated by these length-based tools when a patient’s clinical status precludes obtaining an accurate weight.¹¹⁴ Additionally, international populations, depending on their development status, also vary from the length-based weight predictions set forth by the Broselow–Luten system.¹²⁵