The need for more and better quality data on medication safety in human pregnancy has been in the public and regulatory spotlight for decades. The issue is becoming even more urgent with the impending revisions to the product label format that will eliminate the traditional A, B, C, D, X pregnancy categories, and substitute more data-driven narratives.
In response to this need, over the last several years there has been a steady increase in the number of regulatory requests/requirements for postmarketing surveillance studies for new medications likely to be used by women of reproductive age or by pregnant women. These requests most often have been fulfilled by the initiation of pregnancy registries. However, concern has been raised that pregnancy registries by and large have been challenged to recruit sufficient numbers of study subjects, and the time to completion of these studies is typically longer than desirable. Other questions have been raised about the adequacy of pregnancy registries alone (especially those with small sample sizes and no internal comparison groups) to test hypotheses related to birth outcomes.
In May 2014, the Food and Drug Administration hosted a public meeting to highlight some of these issues and to seek advice and solutions, titled "Study Approaches and Methods to Evaluate the Safety of Drugs and Biological Products During Pregnancy in the Post-Approval Setting." Panelists included representation from federal agencies including the Centers for Disease Control and Prevention, Department of Defense Naval Health Research Center, the Agency for Healthcare Research and Quality, and the Food and Drug Administration. Other panelists represented academic and HMO-based research networks, contract research organizations, the pharmaceutical industry, obstetric providers, and patients. Comments also were provided by audience members who represented a variety of entities including academic societies, professional associations, and advocacy groups.
Specific strategies for obtaining better data in a more timely fashion were presented by the panelists. These included the value and efficiency of utilizing large administrative claims databases or systemwide electronic capture of pregnancy exposure and outcome data. Using this approach, representative samples of patients can be accrued while not requiring individual active informed consent. Limitations of such sources of data also were mentioned, including inability to validate that the pregnant woman actually took a medication of interest, when, and at what dose, and the usual absence of information in claims data on some important confounders such as alcohol use and folic acid supplementation.
Panelists representing various pregnancy registry study designs described approaches to addressing some of the limitations of these studies. For example, some panelists emphasized that "disease-based" registries that involve examination of birth outcomes for several medications used for the treatment of one or more similar maternal conditions have several distinct advantages. These include a more streamlined referral process, whereby clinicians can more easily identify and refer all pregnant patients who have the same underlying condition, irrespective of treatment with any specific medication under study. Examples of successful disease-based approaches that were highlighted by panelists included the Antiretroviral Drugs in Pregnancy Registry, the North American Antiepileptic Drug Pregnancy Registry, and the OTIS/MotherToBaby Autoimmune Diseases in Pregnancy Project. Each of these studies allows for comparison of outcomes across various treatments, while accounting for the possible contribution of the underlying maternal condition to adverse pregnancy outcomes.
Panelists also emphasized that more effective methods are needed for raising awareness of the existence and value of pregnancy registries for providers and consumers alike, including more efficient and extensive use of social media. Panelists, and in particular obstetric providers, indicated a need for better methods of identifying women who are eligible for participation in a pregnancy study, and facilitating the referral process. The obstetric provider panelist suggested that existing electronic medical records systems could be adapted to generate automated alerts to providers regarding patients who qualify for referral to a registry.
The patient representative panelist emphasized the need to engage the pregnant woman in the research process. This was confirmed by research groups whose primary interaction in pregnancy studies is with the pregnant woman herself, resulting in minimal loss-to-follow-up and high participant satisfaction.
Last, there was extensive discussion about the use of other alternative study designs, such as case-control studies, that could help address the limitations of pregnancy registries, especially in terms of statistical power for evaluating rare outcomes such as specific birth defects. The Vaccines and Medications in Pregnancy Surveillance System (VAMPSS) was described as one such alternative, combining the benefits of a cohort (registry-type) study with a concurrent case-control study to address the same exposures in pregnancy.