The 50-year Quest for Better Pregnancy Data
January 3, 2017
The following post was first published in OB/GYN News. Please see our OB/GYN News archives here.
Publish date: November 1, 2016
By: Gerald G. Briggs, BPharm, FCCP, Christina Chambers, PhD, MPH, Lee S. Cohen, MD, & Gideon Koren, MD
Editor’s note: As Ob.Gyn. News celebrates its 50th anniversary, we wanted to know how far the medical community has come in identifying and mitigating drug risks during pregnancy and in the postpartum period. In this article, our four expert columnists share their experiences trying to find and interpret critical pregnancy data, as well as how they weigh the potential risks and benefits for their patients.
The search for information
The biggest advance in the past 50 years is the availability of information, even though limited, relating to the effects of drugs in pregnancy and lactation. In the first few years of this period, it was a daunting task to obtain this information. I can recall spending hours in the hospital’s medical library going through huge volumes of Index Medicus to obtain references that the library could order for me. The appearance of Thomas H. Shepard’s first edition (Catalog of Teratogenic Agents) in 1973 was a step forward and in 1977, O.P. Heinonen and colleagues’ book (Birth Defects and Drugs in Pregnancy) was helpful.
My first edition (Briggs et al., Drugs in Pregnancy and Lactation) came out in 1983 and was followed in 1993 by James L. Schardein’s book (Chemically Induced Birth Defects). In 2001, the first edition of the European book by Christof Schaefer et al. (Drugs During Pregnancy and Lactation) was released.
Although all of the above sources were helpful, any book in an evolving field will not have the newest information. Two important services, TERIS and Reprotox, were started to allow clinicians to contact them for up-to-date data. Nevertheless, the biggest change was the availability of current information from the U.S. National Library of Medicine via Toxnet, PubMed, and LactMed, relating to the effects of drugs in pregnancy and lactation.
The biggest unanswered question is why so many drugs have minimal, if any, human pregnancy and breastfeeding data? In my 11th edition (in press), about 1,443 drugs are reviewed. The majority have little or no human pregnancy data. The situation is even worse for breastfeeding data. In either situation, it places the clinician in a difficult position. How do we inform the patient regarding the potential embryo, fetal, or nursing infant risk? If the maternal benefit from the drug clearly outweighs the unknown risk, then the clinician can explain this to the patient. However, such situations appear to be infrequent and, in breastfeeding, the infant can be bottle fed. In contrast, in most pregnancy cases the comparison of the maternal benefit to the potential embryo/fetal risk is unknown. So what does the clinician do?
My method is to ask three questions. First, are there other drugs with a similar mechanism of action that have some human data? In most cases, the answer to this question is no, but even when there are data, it is typically very limited. Second, does the drug cross the human placenta? The answer is typically based on the molecular weight. Any drug with a molecular weight less than 1,000 daltons probably crosses. In the second half of pregnancy, especially in the third trimester, almost every drug crosses. Third, do the animal pregnancy data predict embryo/fetal risk? It was thought that it could if the dose causing harm was less than or equal to 10 times the human dose based on BSA or AUC and there were no signs of maternal toxicity. However, using data from my 10th edition, I and eight coauthors, all of whom are knowledgeable on the effects of drugs in pregnancy, found that the animal data for 311 drugs raised the possibility of human embryo-fetal harm that current data confirmed in only 75 (24%) of the drugs (Am J Obstet Gynecol. 2015 Dec;213[6]:810-5).
The system needs to be fixed. One method is to give the Food and Drug Administration the authority to require manufacturers of drugs likely to be used in pregnancy to gather and publish data on their use in pregnancy. That sounds reasonable, but will it ever occur?
Gerald G. Briggs
Mr. Briggs is clinical professor of pharmacy at the University of California, San Francisco, and adjunct professor of pharmacy at the University of Southern California, Los Angeles, and Washington State University, Spokane. He is coauthor of “Drugs in Pregnancy and Lactation,” and coeditor of “Diseases, Complications, and Drug Therapy in Obstetrics.” He has no relevant financial disclosures.
Learning the lessons of the past
During the last 50 years, two of the most potent known human teratogens, thalidomide and isotretinoin, became available for prescription in the United States. Thanks to the efforts of Frances Kelsey, MD, PhD, at the FDA, the initial application for approval of thalidomide in the United States was denied in the early 1960s. Subsequently, based on evidence from other countries where thalidomide was marketed that the drug can cause a pattern of serious birth defects, a very strict pregnancy prevention program was implemented when the drug was finally approved in the United States in 2006.
This has produced excellent results in terms of limiting the number of exposed pregnancies. In contrast, when isotretinoin was first approved in the United States in 1982, no pregnancy prevention program was in place. By the late 1980s, it was clear that the drug was linked with a pattern of serious birth defects in numerous infants. Successively more stringent pregnancy prevention/restricted distribution programs were implemented over the years, culminating with the current iPledge program. Despite these unprecedented measures, exposed pregnancies continue to occur and remain a challenge in terms of prevention.
Over the last 50 years, we have also seen an important evolution in our ability to conduct pregnancy exposure safety studies. Though we still have limited ability to conduct clinical trials in pregnant women, the need for good quality observational studies has become more widely accepted. The Centers for Disease Control and Prevention’s National Birth Defects Prevention Study (now in its most recent iteration known as BD STEPS) has been one very important source of data on the safety of a wide variety of medications. Using a case-control study design, women who have delivered infants with specific birth defects and comparison women who have delivered non-malformed infants are interviewed about their exposures in pregnancy. These data have been extremely helpful in generating new hypotheses, confirming or refuting findings from other studies, and in testing hypotheses regarding the safety of medications widely used in women of reproductive age. These analyses, for example, have contributed to the large body of literature now available on the safety of antidepressant medications in pregnancy.
At the same time, in the last 30 years, we have seen a tremendous increase in the number of pregnancy registries required or recommended upon approval of a new drug in the United States. These registry studies, while challenging to complete in a timely manner, have steadily improved in terms of rigor, and several disease-based pregnancy exposure studies have been implemented, which have allowed us to better understand the comparative risks or safety of anticonvulsants and antiretroviral drugs, to name a few.
It is important to note that with all these advances in the last 50 years, we still have a huge gap in knowledge about medication safety in pregnancy and lactation. Recent reviews suggest that more than 80% of drugs currently marketed have insufficient or no data available. If we include over-the-counter medications, the knowledge gap grows larger. With the 2014 approval of the long-awaited Pregnancy and Lactation Labeling Rule, clinicians are now beginning to experience the elimination of the old A-B-C-D-X category system for pregnancy safety. In its place, data-driven product labels are required. These are expected to provide the clinician with a clear summary of the relevant studies for a given medication, and to place these in the context of the background risks for the underlying maternal disease being treated, as well as the population risks. However, it is painfully clear that we have a long way to go to generate the needed, high-quality data, to populate those labels.
Dr. Christina D. Chambers
Dr. Chambers is a professor of pediatrics and director of clinical research at Rady Children’s Hospital, San Diego, and associate director of the Clinical and Translational Research Institute at the University of California, San Diego. She is director of MotherToBaby California, a past president of the Organization of Teratology Information Specialists, and past president of the Teratology Society. She has no relevant financial disclosures.
Moving toward personalized medicine
Nowhere is a lack of actionable data more pronounced than in the impact of mental health drugs in pregnancy.
As Dr. Briggs and Dr. Chambers have outlined, the quality of data regarding the reproductive safety of medications across the therapeutic spectrum has historically been fair at best. The methodology and the rigor has been sparse and to a large extent, in psychiatry, we were only able to look for signals of concern. Prior to the late 1980s and early 1990s, there was little to guide clinicians on the safety of even very commonly used psychiatric medications during pregnancy. The health implications for women of reproductive age are extraordinary and yet that urgency was not matched by the level of investigation until more recently.
The last 25 years, however, has brought a commitment to the systematic study of reproductive safety. From cohort studies, to large analyses of administrative databases and global registries, we’re now able to look at what medications women took during pregnancy and study the specific outcomes. These studies don’t provide complete information, but the dedication to and the identification of reproductive safety as a major public health issue for women is a huge advance.
In psychiatry, we have rapidly improving data informing women about the risk for major congenital malformations. The clinical dilemma of weighing the necessity to stay on a medication to prevent relapse of a psychiatric disorder with the potential risk of malformation in the fetus is a wrenching one for the mother-to-be. Only good information can help patients, together with their physician, make collaborative decisions that make sense for them. Given the same information and the same severity of illness, women will make different decisions, and that’s a good thing. The calculus couples use to make these private decisions is unique to those involved. But they are able to move through the process because they have a platform of high-quality information.
So where do we go in the future? We need to get beyond the question of risk of major malformations and move toward understanding the long-term neurodevelopmental implications of prenatal exposures – whether such exposures confer risk or are even potentially salutary. One needs only look at the vast body of literature regarding fetal exposure to selective serotonin reuptake inhibitors (SSRIs) to observe the realization of this trend. When it comes to SSRIs, a fairly clear picture has emerged that they pose little absolute risk in terms of congenital malformations. What is missing is how SSRIs impact a child’s learning and development at age 3, 5, and 10. There have been a few studies in this area, but not a single, large prospective study that accurately quantifies both exposure to SSRIs and maternal psychiatric illness during pregnancy.
I expect that the future will also bring a greater understanding of the impact of untreated mental illness on the risk for obstetrical, neonatal, and longer-term neurodevelopmental outcomes. Most of the safety concerns have centered around the effect of fetal exposure to medications, but we also need to better understand how untreated psychiatric disorders impact the spectrum of relevant outcomes.
Getting back to the dilemma faced by pregnant women who really need medication to sustain emotional well-being, there simply is no perfect answer. No decision is perfect or risk free. What we can hope is that we’ll have personalized approaches that take into account the best available data and the patient’s individual situation and wishes. We’ve already come a long way toward meeting that goal, and I’m optimistic about where we’re going.
Dr. Lee S. Cohen
Dr. Cohen is the director of the Center for Women’s Mental Health at Massachusetts General Hospital in Boston, which provides information resources and conducts clinical care and research in reproductive mental health. He has been a consultant to manufacturers of psychiatric medications.
Perception of risk
Every year, numerous new medicines are approved by the FDA without data in pregnancy. Animal studies may show a problem that doesn’t appear in humans, or as was the case with thalidomide, the problem may not be apparent in animals and show up later in humans. There are many drugs that are safe in pregnancy, but women are understandably afraid of the potential impact on fetal development.
While my colleagues have presented the advances we’ve made in understanding the actual risks of medications during the prenatal period, it’s also important to focus on the perception of risk and to recognize that the reality and the perception can be vastly different.
I started to look at this area in 1989, just a few years after starting the Motherisk Program. We discovered that women exposed to nonteratogenic drugs were assigning a 25% likelihood of having a malformed child, when the actual risk for those drugs was between 1% and 3% and is unrelated to the drug being taken.
At the same time, we began to ask women, using a visual analog scale, what would be their trend toward continuing or terminating pregnancy? Over several studies, we found that the likelihood of termination was high, and certainly much higher than was supported by the evidence of actual harm to the fetus. Specifically, if a woman received information about the safety of the drug and she still gave more than a 50% probability of terminating the pregnancy when surveyed, there was a good chance that she would terminate the pregnancy.
When you consider that most of the drugs that women are commonly prescribed in pregnancy – from most painkillers to antidepressants – are not known to cause malformations in pregnancy, you begin to see how problematic an inflated perception of risk can become.
But we see different trends in women with serious and chronic health problems, such as lupus or epilepsy. These women are typically under the care of a subspecialist, who in many cases has developed a significant knowledge base and comfort level around prescribing the drugs in this area and is able to communicate more clearly to patients both the risks to the fetus and the consequences of failure to treat their condition.
So clearly, the role of the physician and the ob.gyn. in particular is critical. It’s no secret that physicians face a negative legal climate that encourages defensive medicine and that they are often hesitant to tell women, without reservation, that it is okay to take a drug. But we must all remember that it is very easy to cause a woman not to take a medication in pregnancy and often that’s not what’s best for her health. Many women now postpone the age of starting a family and more have chronic conditions that require treatment. The idea of not treating certain conditions for the length of a pregnancy is not always a viable option. Yet there are quite a few women who would consider termination “just to be on the safe side.” That must be taken very seriously by the medical profession.
Dr. Gideon Koren
Dr. Koren is a professor of physiology/pharmacology at Western University, London, Ont., and a professor of medicine at Tel Aviv University. He is the founder of the Motherisk Program. He reported being a paid consultant for Duchesnay and Novartis.
