Pregnancy registries are valuable sources of information, and for many drugs and vaccines they are the primary source of human pregnancy experience. The strengths of these registries are their prospective nature—women are enrolled before the outcome is known—and enrollment is over a wide geographical area. Typically, two types of pregnancy outcomes are obtained: outcomes with birth defects and outcomes without known birth defects. The latter comprises live births, fetal deaths, and spontaneous abortions.
Registries can identify early signals of teratogenicity, but they have several limitations. They depend on voluntary reporting, which results in selection bias, and they are not representative of target populations. Pregnancies that are lost to follow-up may have had different outcomes than those with documented outcomes. Furthermore, registries lack details on elective terminations and fetal deaths without birth defects, and all spontaneous abortions. Finally, with some exceptions, they usually lack control groups.
Because the total number of exposed pregnancies is unknown, data from a registry cannot be used to calculate prevalence of an outcome, but the data can be used to estimate the proportion of birth defects. Some registries also collect data on retrospective reports, which are less representative of the target population because they can be biased toward the reporting of more unusual and severe outcomes. However, they may be helpful in detecting unusual patterns of birth defects.
A complete list of pregnancy registries is available on the Food and Drug Administration Web site, which provides additional details on the registries, such as fax numbers, links to other Web sites, and mailing addresses (www.fda.gov/womens/registries
Because the strength of a registry is based on numbers, I encourage health care professionals to enroll appropriate patients in these registries whenever possible.