Labeling medications for use in pregnancy

Anthony R. Scialli, BScBiol, MD
Associate Clinical Professor in Obstetrics and Gynecology, George Washington University, Associate Professor in Obstetrics and Gynecology, Georgetown University, Active Staff, Columbia Hospital for Women and Georgetown University Hospital, Director, Reproductive Toxicology Center, Washington, DC, U.S.A.

Among the goals of the product label for drugs is to educate the clinician and the patient, that is, to communicate information that leads to optimal prescribing practice and information that permits an appropriate assessment of pregnancy risk. These goals are sometimes viewed as being in conflict because some practitioners believe that medication exposure during pregnancy always entails a risk of harm to the embryo or foetus; however, failure to prescribe appropriate medication therapy may have adverse consequences for the mother and for the conceptus and these adverse consequences may be more severe than the actual or perceived medication risk to the unborn.

A medication label should ideally include the following:

Whether and to what extent such information can be included on the product label will depend on information availability, to be sure, but also on the interests of the manufacturer and on the traditions of the relevant regulatory body, e.g., the Food and Drug Administration (FDA) in the U.S. For example, it is unlikely that the label will explore a comparison of available alternative treatments for use during pregnancy, particularly when the alternatives are marketed by different companies. Furthermore, the available space is limited.

A common situation arises for many pharmaceutical agents, particularly newly marketed ones, for which the estimation of developmental toxicity risk must be made largely on preclinical data. There is considerable disagreement among clinicians regarding the interpretation, and even the interpretability, of experimental animal studies; however, it is not useful for the label to remain mute on the issue of risk. Nonspecific advice such as "risk should be balanced against benefit" is of no assistance; rather, it will be necessary to develop methods of communicating to clinicians the meaning of preclinical study results. In addition, communication of the results of human studies should include the limitations of these studies because some clinicians will believe that any human data, however poor, are superior to any experimental animal data, however good.

The material in the label should be explicitly stated, without use of code words or terms charged with emotional content. Risks should not be characterized as high or low, but rather should be stated numerically when possible and in the context of the baseline risk of adverse outcome in the general population. The use of focus groups of practitioners and of non-practitioner lay people will be helpful in identifying language that is confusing or inappropriately alarming.

In summary, the ideal label should use permissive language to give the practitioner and the patient explicit approval to use the therapy, i.e., that it is good to use in pregnancy if you need to treat a particular disease or disorder; and it should inform on therapeutic choices and advise on inadvertent exposures, be informative and impartial.


  1. Code of Federal Regulations (United States), Title 21 (Food and Drugs), Vol 4, Part 201, Labeling [CITE: 21CFR201.57]. Revised as of April 1, 1998:22-4.
  2. See also: for a display of proposed new labeling guidelines.

Back to home