No doubt most pregnant patients want clear-cut answers to that question, but as this primer on teratology points out, sometimes the answer is straightforward, and sometimes it's not.
What are the odds of a pregnant patient being on a prescription medication? One recent French study indicated that 99 out of 100 patients had filled at least one script during their gestation, with almost half being written by family practitioners (49%).1 On average, each woman in this survey took 13.6 medications. When you add these medications to the long list of nonprescription drugs and environmental chemicals that the average woman is exposed to, it's obvious that quite a few women are exposed to potential teratogens.
Of course, drug use is common among women of childbearing age even when they're not pregnant, but perhaps because of the physiologic changes associated with pregnancy and symptoms often associated with those changes, drug consumption may increase during pregnancy.2 Research in past decades would certainly suggest that: In 1970, pregnant women in the United States on average took 8.7 medications.3 In 1973, it had increased to 10.3, and by 1978, to 11.4,5
How do all these drugs affect the developing fetus? It's a little hard to fathom the fact that prior to 1941, few experts believed that environmental factors could adversely affect the human embryo. But this naiveté was shattered by Gregg's observations that maternal rubella infection caused serious harm to the fetus.6 And today, we realize that such teratogens not only contribute to congenital anatomic malformations but have short- and long-term functional effects on the fetus as well. Our purpose here is to define the principles of teratogenesis, discuss the primary effects of teratogens, and offer some practical advice on how a knowledge of these principles can improve day-to-day patient care.
A fetus exposed to a teratogenic agent during prenatal life can experience a variety of postnatal changes in morphology or function, changes that are usually permanent. Teratogens can be chemicals, drugs, infectious agents, or physical changes. We'll confine most of the discussion to prescription drugs. Whatever the agent, after maternal exposure, it must cross the placenta to reach the embryo or fetus. Exceptions to the rule are agents that are directly administered to the fetus, which are beyond the scope of this article.
The teratologic potential of a drug depends on a variety of factors: The physicochemical characteristics of the drug, maternal and fetal blood flow to and from the placenta, and the physiology of the placenta itself. The misconception that the placenta is a barrier to the transfer of drugs between mother and fetus has long been discarded; instead recent research indicates that the placenta undergoes continuous structural and functional changes during gestation that can affect the rate and extent of drug transport. Unfortunately, while animal studies have shown that drug transport varies during gestation, there is little information on the effect of placental differentiation and aging on drug transfer in normal human pregnancy; nor is there a great deal of data on placental drug transport during various diseases.
We don't fully understand how certain drugs contribute to fetal anatomic malformations, but estimates suggest that they are responsible for up to 5% of malformations. Of greater importance is the fact that teratogens, often at far lower doses than those needed to produce anatomic anomalies, can induce subtle defects in the biochemistry, physiology, and behavior of the developing fetus.7 In addition, many of these subtle defects are not apparent at birth, and may remain unrecognized until years later. Like classic teratogens, which produce anatomic malformations, these drugs do so at critical periods of fetal development. Their effect occurs beyond the period of organogenesis.
One well-known example is DES (diethylstilbestrol), which causes adenocarcinoma of the vagina after puberty with in utero exposure. Another example is the effects of androgens on females exposed in utero. If androgen exposure is high enough and if it occurs during the critical period of fetal brain development, females will exhibit masculine behavior after birth.8
Wilson and Fraser developed the general principles of teratology following the devastating effects of thalidomide in the early 1960s.9 These principles state that the effect of a drug as a teratogen is dependent upon: (a) gestational age at the time of exposure; (b) the duration of exposure; (c) the genotypes of the mother and fetus; and (d) the dose reaching the embryo or fetus. Let's take a more detailed look at these basic principles.
Timing. The timing of exposure is the most critical factor for determining fetal risk. For most structural defects, exposure must occur during organogenesis. In humans, this period is 20 to 55 days after fertilization (3570 days after the first day of the last menstrual period). Within this period, different organs and structures have different periods of sensitivity, during which their development can be disrupted.
In general, drug exposure that occurs more than 10 weeks after the last menstrual period will not cause anatomic congenital malformation. There are, however, several exceptions to this rule, including alcohol, which can cause brain defects when exposure occurs anytime during gestation, and several drugs that produce malformations if exposure occurs during the second and third trimester. The list includes tetracycline, which causes teeth staining, iodine (goiter), angiotensin-converting enzyme (ACE) inhibitors (hypocalvaria, renal defects), and angiotensin II receptor antagonists, which cause the same defects as ACE inhibitors.
Genetic factors. Maternal and fetal genetic characteristics have a major influence on the ability of a specific agent to cause structural birth defects. Although the cause of most structural anomalies is unknown, the majority of known causes are genetic. These defects can result from mutations in a single gene or by a combination of genetic and environmental factors. Achondroplasia, for instance, is a monogenetic defect whereas fetal hydantoin syndrome, which is characterized by absent or low levels of epoxide hydrolase, belongs to that multifactorial group.
Environmental exposures, on the other hand, including drugs, radiation, hypothermia, and chemicals, are a relatively uncommon cause of congenital malformations, accounting for less than 5% of all birth defects.10 Congenital malformations overall occur in up to 3% of all births.
Dose. The maternal dose, or more specifically the dose reaching the embryo or fetus, plays a major role in determining an agent's teratologic potential. In general, the dose required to cause embryo/fetal dysfunction is less than or equal to the dose that causes intrauterine growth restriction (IUGR), which is less than the dose that causes malformation, which in turn is lower than the lethal dose. In addition, acute or short-term dosing presents a greater risk than chronic dosing. This may be the case because chronic exposure can induce drug- metabolizing enzymes that somewhat mitigate the agent's toxic effects.10
Of course, federal regulatory agencies require that a dose-response relationship be determined in animal reproductive tests. These teratology studies are able to look at a wide range of doses, but to be clinically meaningful, the animal dose eliciting a specific response should be stated in terms of body surface area and compared to the maximum recommended human dose expressed in the same units. If systemic concentrations are measured to determine the dose-response curve related to the drug's teratogenic effects, this should be reported as area under the plasma concentration versus time curve (AUC). (In an x/y graph that plots dose response against time, AUC is the range of dosages that falls below the line that represents the graph's curve.) Because AUC is a measure of plasma drug concentration over time, it may also be a measure of how much drug is available to cross the placenta.
Obviously, it's impossible to test potential teratogens in humans using the same wide range of doses used in animals. Because of the narrow therapeutic range of doses used clinically, a dose-response relationship in humans is not available, with some exceptions, including cigarette smoking, alcohol, ionizing radiation, vitamin A, and intra-amniotic methylene blue.11
Knowing the teratogenic dose in animals may have some value in trying to predict an agent's effect on humans. Keep in mind, however, that animal data will only apply clinically if the teratogenic dose does not also cause maternal toxicity.
In general, doses causing teratogenic effects in animals that exceed the maximum recommended human dose by ten- to hundredfold pose less of a threat to humans when compared to an animal teratogenic dose that is less than 10 times the maximum recommended human dose. At this threshold, the potential for human teratogenicity becomes a major concern. Drugs with an animal teratogenic dose less than the human dose (e.g., ribavirin) should be considered potential human teratogens until proven otherwise.
Unfortunately these dose comparisons may not apply to women with unique genetic sensitivities or in cases of overdose.9 When very high doses (e.g., those more than 100 times the maximum human dose) are required to produce structural defects in animals, without accompanying maternal toxicity, the outcomes usually cannot be extrapolated to humans.
Seven criteria have been proposed to establish an agent as a human teratogen:12
(a) exposure to the drug at critical times during gestation;
(b) consistent findings by epidemiologic studies;
(c) case reports that suggest teratogenicity;
(d) a rare exposure associated with a rare defect;
(e) the frequency of specific outcomes associated with introduction or withdrawal of the agent;
(f) teratogenicity in animals; and
(g) biological plausibility.
A drug dose doesn't have to meet all of the criteria, but three of the criteria (a, b, and c or a, c, and d) are essential to establish human teratogenicity. Of all these criteria, case reports have played the most prominent role over the years, providing the first evidence of teratogenicity for about three fourths of the drugs that are now considered proven human teratogens.10
An embryo or fetus can be exposed to thousands of drugs and chemicals, but only a few are proven or probable human teratogens when typical doses are administered to the mother at a critical time in gestation (Table 1). There are other agents that are potent teratogens in animals (Table 1), but in these cases, the evidence for human teratogenicity is either inadequate or completely lacking. Unfortunately avoiding exposure to all of these substances during the critical periods of pregnancy may not reduce the number of infants born with a major birth defect. As discussed previously, drugs are not a major cause of birth defects.
As we already mentioned, the incidence of major congenital anomalies recognized at birth is approximately 3%; but the true incidence of major structural defects may be two or three times the number observed at birth. Many major malformations are not appreciated or manifest at birth. It may be months or years before they are recognized and diagnosed as birth defects. This is also true for carcinogenic effects, the classic example being vaginal adenocarcinoma occurring in the second decade of life as a result of in utero exposure to diethylstilbestrol, or neurobehavioral defects that result from fetal exposure to alcohol, which may only be appreciated when the child enters school.
Chambers and associates have developed a unique method to identify congenital malformations in newborns exposed in utero to drugs. They cite evidence to show that most drug-induced congenital malformations are characterized by a pattern of major and minor malformations, rather than a single major defect. Their approach combines traditional methods of surveillance with standardized examinations of children, within 6 months of birth, by pediatricians trained in dysmorphology.13
The principles of teratology can be used to create an algorithm that predicts the degree of risk posed by a specific agent (Figure 1). One of the first questions to be answered is whether or not the drug is a known teratogen. If a new drug is in the same pharmacologic class as known human teratogensfor example, danazol is an androgenor if it has a similar mechanism of action (valsartan is an angiotensin II receptor antagonist, a new pharmacologic class that has a similar mechanism of action as ACE inhibitors), then the new drug is probably a human teratogen. Typically, however, most drugs (new and old) have insufficient human pregnancy experience to say that they represent either a high or low risk to the embryo/fetus. The safest course is to avoid any drug during the critical period (usually the period of organogenesis or the first trimester for a greater margin of safety).
If you must prescribe the drug during this sensitive period, then examine the animal reproduction studies that have been conducted on the drug. The common experimental animal species (mouse, rat, and rabbit) have placentas that are very close to the human placenta so we can assume that the drug concentrations in the animal embryo or fetus are similar to those achieved in humans. In addition, if the routes of administration in animals and humans are the same, then one can assume that the metabolic profiles are the same. For drugs that are teratogenic in animals, the critical factor then becomes the teratogenic dose and its comparison to the human therapeutic dose. As previously mentioned, teratogenic doses that exceed the maximum human dose by less than tenfold suggest a high human risk. This risk diminishes as the dose comparison increases. For the greatest margin of safety, if no animal teratogenicity or maternal toxicity has been observed with a dose 100 times or more than the maximum human dose, then the human risk is probably very low.
While most of our discussion has focused on drugs, the teratologic effects of ethanol have been recognized since antiquity. However, experts have only recently turned their specific attention to fetal alcohol syndrome. The syndrome produces the characteristic pattern of craniofacial anomalies at birth in a woman drinking four to five drinks per day during early pregnancy. As few as two drinks a day during early pregnancy may also be risky because the result may be low birthweight. This effect on birthweight was also noted when the father regularly had two drinks a day (30 mL) of ethanol, at or before the time of conception.
Women who drink during pregnancy can also see behavioral effects in their offspring even in the absence of anatomic defects. Those deficits include poor cognitive development and problems with school performance and attention. The effects of social drinking are more controversial, with both positive and negative results in prospective studies that have been conducted. These long-term effects require further investigation.
We don't really know what a safe level of alcohol consumption is during pregnancy. The American Council on Science and Health recommends that pregnant women limit their alcohol consumption to no more than two drinks daily (1 oz or 30 mL of absolute alcohol). However, the safest course for women who are or are planning to become pregnant is abstinence.
Marijuana and hashish may likewise pose a threat to a pregnant patient. The main psychoactive ingredient in marijuana and hashish is delta-tetrahydrocannabinol, which is also available in a commercial oral formulation (dronabinol). While many pregnant women use these drugs, their effects on the unborn remain quite controversial. That's primarily because of the close association between marijuana, alcohol, nicotine, other drugs of abuse, and lifestyles that may increase fetal and neonatal risk. Most studies have documented longer gestation and a decrease in fetal growth. In utero exposure to marijuana has also been linked with increased neonatal tremors and exaggerated startles, both spontaneous and in response to minimal stimuli. On follow-up examinations, the abnormalities in neonatal behavior apparently did not result in poorer performance on cognitive and motions evaluations at 18 and 24 months. Additional prospective studies are needed to resolve these questions.
In spite of all the warnings, many women continue to use potential teratogens during pregnancy. Fortunately, the vast majority of these women still manage to deliver a healthy newborn. Nonetheless, awareness of the known human teratogens and toxins, the critical periods of exposure, and the types of malformations or toxicity will help clinicians better serve their patients and reduce the threat of malformations.
The risk of drug-induced congenital malformations can be reduced in three ways: (1) Whenever possible, use only those drugs considered to be low risk in pregnancy. (2) If you must prescribe a teratogenic or toxic drug, avoid the critical period associated with its toxicity. (3) Use new drugs with little or no human pregnancy experience only after determining the estimated magnitude of risk.14 Although no therapeutic agent is entirely safe in pregnancy, carefully evaluating each patient's risk improves her odds of a positive outcome. For further information on a wide variety of drugs used in pregnancy, refer to the reference guide: Drugs in Pregnancy and Lactation. 6th ed. 2002. By Briggs GG, Freeman RK, Yaffe SJ. (Lippincott Williams & Wilkins, Philadelphia, Pa.).
1. Lacroix I, Damase-Michel C, Lapeyre-Mestre M, et al. Prescription of drugs during pregnancy in France. Lancet. 2000;356:1735-1736.
2. Brockelbank JC, Ray WA, Federspiel CF, et al. A controlled study of Tennessee Medicaid recipients. Drug prescribing during pregnancy. Am J Obstet Gynecol. 1978;132:235-244.
3. Bleyer WA, Au WY, Lange WA, et al. Studies on the detection of adverse drug reactions in the newborn. I. Fetal exposure to maternal medication. JAMA. 1970;213:2046-2048.
4. Hill RM. Drugs ingested by pregnant women. Clin Pharmacol Ther. 1973;14:654-659.
5. Doering PL, Stewart RB. The extent and character of drug consumption during pregnancy. JAMA. 1978; 239:843-846.
6. Gregg NM. Congenital cataract following German measles in the mother. 1941. Aust N Z J Ophthalmol. 1991;19:267-276.
7. Shapiro BA, Goldman AS. New thought on sexual differentiation of the brain. In: Vallet HL, Porter IH, eds. Genetic Mechanisms of Sexual Development: proceedings of a symposium sponsored by the Birth Defects Institute of the New York State Dept. of Health, and held in Albany, N.Y., Nov. 8-9, 1976. New York, NY: Academic Press; 1979:221-246.
8. Money J, Mathews D. Prenatal exposure to virilizing progestins: an adult follow-up study of 12 women. Arch Sex Behav. 1982;11:73-83.
9. Wilson JG, Fraser FC, eds. Handbook of Teratology. New York, NY: Plenum Press; 1977-1978.
10. Schardein JL. Chemically Induced Birth Defects. 3rd ed. New York, NY: Marcel Dekker, Inc.; 2000.
11. Shepard TH. Dose response in human teratology. Teratology. 2002;65:199-200.
12. Shepard TH. Catalog of Teratogenic Agents. 10th ed. Baltimore, Md: Johns Hopkins University Press; 2001.
13. Chambers CD, Braddock SR, Briggs GG, et al. Postmarketing surveillance for human teratogenicity: a model approach. Teratology. 2001;64:252-261.
14. Briggs GG. Drug effects on the fetus and breast-fed infant. Clin Obstet Gynecol. 2002;45:6-21.
Sumner Yaffe. "Is this drug going to harm my baby? Contemporary Ob/Gyn Nov. 1, 2003;48:57-68.