Dr. Howland is Associate Professor of Psychiatry, University of Pittsburgh School of Medicine, Western Psychiatric Institute and Clinic, Pittsburgh, Pennsylvania.
The author discloses that he has no significant financial interests in any product or class of products discussed directly or indirectly in this activity, including research support.
Address correspondence to Robert H. Howland, MD, Associate Professor of Psychiatry, University of Pittsburgh School of Medicine, Western Psychiatric Institute and Clinic, 3811 O’Hara Street, Pittsburgh, PA 15213; e-mail: HowlandRH@upmc.edu.
The treatment of mental disorders in women of child-bearing age is an important clinical issue (Altshuler et al., 1996). How do we know about the safety of medications during pregnancy and lactation? In this article, the first of three on this topic, I describe the kinds of studies conducted to evaluate safety. Next month, I will describe the current and proposed U.S. Food and Drug Administration (FDA) classification system for pregnancy and lactation labeling. In the final article, I will discuss some general principles and guidelines for prescribing medications during pregnancy and lactation.
Sources of Information About Medication, Pregnancy, and Lactation
Conducting randomized clinical trials (RCTs) comparing one or more drugs and placebo in patients who are pregnant or breast-feeding is considered unethical, because of the potential risk exposure to the fetus or neonate. Drug data on pregnancy outcomes and lactation effects are first derived from animal studies, but such studies cannot guarantee safety in lieu of human research. Human studies include case reports and epidemiology studies. Single or multiple case reports of negative outcomes cannot establish causality, but these reports can be used to generate hypotheses that can be investigated in larger epidemiological studies.
In the absence of RCTs, epidemiology studies are the best way to identify associations between a given drug exposure and abnormalities in the newborn (FDA, 2005b). Demonstrating absolute safety is not possible from epidemiology studies, but the findings can provide reassurance if the risks are not found to be elevated. The ability of a study to demonstrate safety or risk, however, will depend on its sample size (“power”). Large studies have greater “power” to detect even small associations between a drug and birth defects, although smaller studies can show risk, if the risk is of a high magnitude.
Because various mental disorders, medical conditions, nicotine, alcohol, and illicit drugs can affect pregnancy outcome, whether particular medications independently cause birth defects can be confounded by the presence of these other clinical variables. Other characteristics such as the mother’s pregnancy history, age, race, weight, geographical location, and socioeconomic status can affect pregnancy outcomes; they also can be confounding variables in epidemiology studies. Hence, it may be difficult to determine whether an observed risk of increased birth defects is due to a particular medication exposure, to any of these clinical variables, or to an interaction among them. Epidemiology studies can be prospective or retrospective, and each design has strengths and limitations.
Prospective Epidemiology Studies
Prospective studies include cohort studies and pregnancy exposure registries (FDA, 2005b). Cohort studies enroll a group of pregnant women before pregnancy outcome is known and collect relevant clinical and demographic information periodically throughout pregnancy, including medication use. Offspring are then examined at birth and followed for a certain amount of time. The strength of the cohort design is the prospective, systematic collection of data, including drug exposures, confounding variables, and outcome information. If the population is large, the study can examine multiple exposures and multiple outcomes. Generally, however, small numbers of specific birth defects will be seen even in a large study population. For this reason, if only a small number of women are exposed to specific drugs within the cohort, the detection of even a substantial drug-induced increase in the rate of any individual defect will be very difficult. In addition, the absence of an appropriate control group for comparison makes it difficult to definitively establish causality for a drug exposure and pregnancy outcome.
Pregnancy exposure registries are a kind of cohort study that prospectively enroll pregnant women exposed to one or more specific drugs of interest and evaluate the associated pregnancy outcomes. For comparison, non-drug-exposed pregnant women might be included. Detailed assessments that gather data on drug exposures and other confounding variables are conducted periodically throughout pregnancy, and infants are sometimes followed for a period time to assess developmental effects. Adverse pregnancy outcome rates in the drug-exposed cohort are compared with rates in the non-exposed group or from population data. This design is able to evaluate a range of adverse outcomes, but it is likely to undercount spontaneous drug-related abortions if they occur before pregnancy is known. The length of time an infant is monitored (from birth onward) and the source of information about the infant (e.g., from the parent, obstetrician, pediatrician) also can influence the number and kinds of defects detected.
Pregnancy exposure registries are limited by the choice of an appropriate comparison group because the comparison (exposed or not exposed) is not randomized. Comparison to the general population background rate of birth defects can be done, but this does not take into account the influence of all potentially relevant maternal characteristics that might be related to the occurrence of birth defects. Matching the exposed group with a non-exposed comparison group (a group of pregnant women similar in all respects, including the disorder leading to drug exposure, but not taking the specific drug) can be done but is very difficult. Registries can be limited by self-referral bias (i.e., women who enroll may be more or less likely to have infants with defects). Dropouts during follow up also can limit the interpretability of registry data, especially if dropping out is related to whether or not the infant is normal. Pregnancy exposure registries may be too small or too limited to resolve concerns about a drug, increasing the risk of specific birth defects, unless the teratogenic drug effect is so powerful that defects occur in a relatively high proportion of exposed pregnancies.
Postmarketing pregnancy exposure registries are often used proactively to monitor for major fetal effects, and this is the best way to collect information on exposures with newly marketed drugs. Manufacturers may develop pregnancy exposure registries, either on their own initiative or when requested by the FDA as a postmarketing directive. Many lists of current pregnancy exposure registries enrolling women are often available from online searches (FDA, n.d.).
Retrospective Epidemiology Studies
Retrospective epidemiology studies include birth defect registries and case-control studies (FDA, 2005b). Birth defect registries monitor temporal and geographical frequencies of birth defects. The rationale for these registries is that the introduction of a new major teratogen would lead to an unusual frequency or clustering of particular defects. Possible teratogens would include chemicals, infectious agents, drugs, or any other kind of exposure that results in fetal malformations. Birth defect registries usually cannot detect drugs as teratogens, but they have been used to identify cases for use in case-control drug studies and to provide population-based birth defect rates for use as a comparator in pregnancy exposure registries.
Case-control studies enroll infants with specific birth defects and collect data on gestational exposures, including drugs. The frequency of drug exposure in the cases is compared with the frequency of exposure to the same drug in a group of controls without the birth defects of interest. Drug exposure information usually is obtained retrospectively by interviews, questionnaires, or review of medical records.
Although pregnancy exposure registries are limited to screening for major drug teratogens (drugs having common or likely teratogenic effects), case-control studies have the statistical power to identify drug teratogens with rare or infrequent levels of risk. Recall bias is a significant limitation with case-control studies because of limitations to the specificity with which mothers can recall their drug exposures during pregnancy, especially if a drug is used only occasionally or if more than one drug is used. Medical or pharmacy records can be used to confirm drug exposures, verifying that a drug was prescribed or a prescription filled, but they cannot determine whether the drug was actually ingested. A mother of an infant born with a major birth defect also may be sensitized to remember more clearly all gestational events and exposures than a mother of a normal infant, and this kind of recall bias must be taken into account in the design and analysis of case-control studies. Enrolling infants with a wide range of other defects as controls may help correct for this recall bias.
Infants could be adversely affected by drug exposure during breast-feeding, but this will depend on understanding the extent of drug transfer into breast milk, the effects of drugs on milk production and composition, and the extent and consequent effects of exposure to drugs in breast milk on breast-fed infants. Ideally, these studies would involve measures of the drug oral dosage and plasma concentration in the mother, drug concentration in the breast milk, plasma or urine drug concentration in the infant, and clinical assessments of the infant (FDA, 2005a). If the pediatric effects of a drug are known, determining the maternal oral dosage, plasma concentration, and breast milk concentration might be sufficient.
Problems with conducting lactation studies include inter-individual maternal and infant variability. Maternal factors that can affect lactation and factors that affect the pharmacokinetics (i.e., absorption, distribution, metabolism, excretion) of a drug to be studied will vary among women. Infant factors (e.g., age, extent of breast-feeding, and age-related changes in pharmacokinetics) will also vary. Obtaining a large enough sample to investigate lactation effects for a particular drug is also difficult. For these reasons, there are relatively fewer studies conducted investigating the use of medication during lactation, and the results of these studies are more difficult to generalize.
Knowing the effects of medication during pregnancy and lactation is a critically important issue, but our ability to obtain such data is limited in certain ways. Different studies sometimes come to different conclusions. These differences are likely explained by the particular design of the study. Nurses who see and counsel patients about pregnancy and breast-feeding should understand the kinds of studies and their limitations as sources of information about medication use, pregnancy, and lactation.
- Altshuler, LL, Cohen, L, Szuba, MP, Burt, VK, Gitlin, M & Mintz, J. 1996. Pharmacologic management of psychiatric illness during pregnancy: Dilemmas and guidelines. American Journal of Psychiatry, 153, 592–606.
- U.S. Food and Drug Administration. 2005a. Guidance for industry. Clinical lactation studies—Study design, data analysis, and recommendations for labeling. Retrieved January 16, 2009, from http://www.fda.gov/cder/guidance/5918dft.htm
- U.S. Food and Drug Administration. 2005b. Reviewer guidance. Evaluating the risks of drug exposure in human pregnancies. Retrieved January 16, 2009, from http://www.fda.gov/cder/guidance/6777fnl.htm
- U.S. Food and Drug Administration. (n.d.). Pregnancy. Retrieved January 16, 2009, from http://www.fda.gov/womens/healthinformation/pregnancy.html