of denominator data on number of users to delineate the frequency of an event, lack of control groups, recall bias of patients and reporters, poor case documentation in the reports (critical details that could contribute to an understanding of an event are missing), and substantial underreporting of AEs (Ahmad et al., 2005).
Although AERS data may provide the initial signal of a safety problem, other studies and databases are typically needed to investigate associations. Those data include results of clinical trials and epidemiologic studies that are conducted, or whose results are available, after a drug is approved.
The sponsors’ phase 4 trials are intended to expand the understanding of the safety and efficacy profile, of selected drugs. However, many of these studies are not completed (or even begun), for various reasons described above. FDA lacks the regulatory tools to adequately compel sponsors to complete appropriate studies (see Chapter 5 for more information). According to a March 2006 report, out of 1,231 agreed-on (by the sponsor) open postmarket commitments of drugs and biologics, 797 (65 percent) have yet to be started12 (FDA, 2006d).
Another source of postmarket safety data is studies of marketed drugs designed to investigate new or expanded indications. Sponsors may include these studies in an efficacy supplement submitted to FDA seeking expanded label indications. Sometimes these studies may yield important data. For example, the APPROVe (Adenomatous Polyp Prevention on Vioxx) trial was designed to identify a new application for rofecoxib and showed an increased risk of serious cardiovascular events with rofecoxib compared with placebo—this cardiovascular impact was a secondary consideration (FDA, 2004b). Post-marketing safety information may also be generated by sponsors through the establishment of active surveillance systems, such as pregnancy-exposure registries (Ackermann Shiff et al., 2006).
The National Institutes of Health (NIH) or other agencies may also sponsor trials to gain new information about marketed drugs. Examples are the NIH-funded randomized controlled primary prevention trial, the Women’s Health Initiative, which reported on adverse health effects of and benefits from use of combined estrogen and progestin (Rossouw et al., 2002). An earlier NIH-funded study, the cardiac arrhythmia suppression trial, found that drug treatment for asymptomatic ventricular arrhythmia in patients who had a heart attack did not prevent—and in fact substantially increased the risk of—sudden cardiac death. FDA had used a drug’s effect on