the adverse events that were being reported for Halcion might be due, at least in part, to the use of Halcion for much longer periods of time and at higher doses than those currently recommended in the labeling.1

The committee believes that this type of use may be a problem common to all hypnotic medications and that it is complicated by incomplete understanding of insomnia and its clinical management. Although prescription of hypnotic drugs at higher doses and for longer durations than those recommended in the product labeling may provide benefit to some patients, the magnitude of Halcion use at higher doses and for longer durations than those that are recommended, also suggests that alternatives (e.g., other medications or diagnoses) are not fully explored, to the potential detriment of patients.

Spontaneous reporting of adverse events provides a "signal" to FDA of the possibility of serious unintended threats to the health of the patient. The pharmacoepidemiolgist, among others, then has the task of deciding which signals should be followed up and which can be ignored. The severity of the events, the size of the at-risk population (and the potential for larger numbers of adverse events), and information concerning use at higher doses or for longer durations than those that are recommended are all important factors in the decision to pursue the spontaneous report(s) further. Further investigations, if they are conducted, could include the following:

  • Verification of possible adverse drug reactions (ADRs);

  • Collection of estimates of drug use in a population;

  • Search for more ADRs attributable to the suspect drug;

  • Examination of in toxicology animal regarding the suspect drug;

  • Examination and reanalyses of the data from clinical trials;

  • Launching ad hoc case-control or cohort studies exploring the association of the drug and the suspected adverse event;

  • Querying various drug surveillance systems under contract to FDA and regulatory bodies in other countries;

  • Querying the drag company that markets the suspect drug; and

  • Research studies with other designs, including rechallenge and withdrawing the drug (experiment in prevention).

Postmarketing surveillance requires the collection and assessment of at least two very different types of information: data from controlled trials, and data from spontaneous reports of adverse events. These two types of data vary significantly in their quality, and, thus, their interpretation as a body Can be quite complicated. This was true for Halcion, because some of the clinically significant adverse events (e.g., memory impairment, nervousness) were detected not in the clinical trials but only in the spontaneous reports. In such circumstances and in those instances in which adverse events are difficult to detect—but are clinically


The FDA task force also observed that "marketing data suggest that Halcion is sometimes prescribed by physicians for longer periods of time and at higher doses than is recommended in the labeling" (FDA, 1996, p. iii).

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement