Expecting the Unexpected — Drug Safety, Pharmacovigilance, and the Prepared Mind
http://www.100md.com
《新英格兰医药杂志》
When a new drug is first marketed, findings regarding its efficacy and safety are commonly based on the experience of several thousand people who have been treated in controlled clinical trials. Despite extensive testing, rare adverse events (those that occur in less than one patient per thousand) can easily escape detection, and unforeseen interactions with coexisting clinical conditions or other drug therapies may remain unexplored. As a result, the characterization of the full safety profile of a new drug relies heavily on clinicians' careful observation of its effects in "real world" practice that is far removed from clinical-trial conditions. Discovery in an observational science such as pharmacovigilance depends on the capacity to recognize and investigate unexpected clinical events that are manifest once a new drug is in use. The detection of such unanticipated effects hinges on what Pasteur called "the prepared mind."
In this issue of the Journal, Bennett and colleagues (pages 1403–1408) demonstrate the power and importance of the detection, analysis, and remediation of adverse drug events. When there was a sudden increase in the rate of sporadic reports of pure red-cell aplasia in association with the use of epoetin, it prompted an analysis of the worldwide adverse-event experience of pure red-cell aplasia in patients treated with epoetin. Careful analysis of the multinational clinical experience of many patients, in turn, suggested that subcutaneous administration and variable storage conditions are risk factors for epoetin-associated pure red-cell aplasia. Interventions designed to minimize these risk factors were followed by an 83 percent decrease in the number of reported cases.
Underlying the successful elucidation of this association was the recognition of a potential link between the drug and the adverse event of pure red-cell aplasia. In this specific instance and in general, recognition is aided by the degree of unexpectedness of the event, given the circumstances of the individual patient, the underlying disease, and background rates of the particular type of event. Highly unusual or infrequent outcomes (such as pure red-cell aplasia) are strong triggers of suspicion about the possible contributing role of a drug. Similarly, awareness may be heightened when an unexpected and otherwise unexplained adverse outcome is often drug-related; examples include Stevens–Johnson syndrome, bone marrow or liver failure, and torsades de pointes. If the adverse event in question occurs at a high background rate or may arise from the underlying disease, the recognition of an association with a drug is more difficult and may hinge on a finding of a strong temporal relationship between the event and drug therapy, abatement of the adverse effects when therapy is discontinued, or recurrence on reexposure to the drug. These and other factors that suggest a possible drug-related adverse event are listed in the Table.
Table. Factors Suggestive of a Possible Adverse Drug Reaction.
Once the potential linkage of an adverse event with a drug has been recognized, pertinent clinical information must be shared with a wide audience of other clinicians. The sharing and assessment of such information can occur through professional meetings, in the scientific literature, or through the reporting of adverse events to the relevant pharmaceutical company or one of its representatives. In the United States, Japan, and many other countries, including those of the European Union, drug companies must report to the appropriate regulatory authorities any adverse events that are made known to any company representative. In the United States, the public health and regulatory jurisdiction of the Food and Drug Administration (FDA) over pharmaceutical manufacturers frees clinicians from constraints with regard to the disclosure of protected health information without a patient's permission; the reporting of adverse events and postmarketing safety surveillance are explicitly exempted under the Health Insurance Portability and Accountability Act (HIPAA).
In the case of epoetin and pure red-cell aplasia, the compilation and analysis of the data on adverse events revealed risk factors that may have contributed to the development of the disease. The FDA, the World Health Organization (WHO), and other regulatory authorities maintain databases of adverse-event reports and analyze them systematically for new safety signals. The FDA's Adverse Event Reporting System, sometimes called MedWatch for the reporting form (www.fda.gov/medwatch/safety/3500.pdf) and outreach program encouraging reporting on the safety of medical products, has collected information since 1969 and now contains more than 2.5 million adverse-event reports for several thousand drug products. Within the FDA, staff members in the Office of Drug Safety review all individual case reports involving outcomes that are life-threatening, that result in death, that lead to hospitalization, disability, or congenital anomalies, or that require interventions to prevent clinically significant impairment. Targeted and comprehensive investigation and analyses may be triggered by one striking case report, by an unusual pattern of adverse events, or by a collection of adverse-event reports exceeding that expected in usual clinical experience or a given clinical context. The FDA, like the WHO and government agencies in other countries, also uses statistical methods of data mining to help discern which combinations of drugs and adverse events in their databases are distinctive. Data mining can trigger additional case-study analyses or population-based studies.
Adverse events reported by clinicians need to be placed in context so that the magnitude of the potential risk associated with a given product can be estimated. The volume of drug use often provides a crude context for the level of risk: reporting rates of adverse events may be derived using the volume of prescriptions written or dispensed as a surrogate for population exposure. Whenever possible, true population-based incidence rates are the preferred method for describing and comparing the levels of risk associated with various products; reporting rates are based only on those adverse events that are actually reported, which may represent as little as 1 percent of all cases. In general, adverse-event reporting is greatest during the first few years of a product's marketing, and reporting can be stimulated by scientific or media attention.
The ability to detect drug-safety signals through spontaneous reports is limited by the low, and highly variable, percentage of events that are reported, the uncertain estimates of exposure represented by the number of prescriptions written, and the passive nature of data collection. There are a few systems that use active ascertainment of adverse events from doctors. In the United Kingdom, reports of such events are actively solicited through the Prescription-Event Monitoring system, which surveys prescribers regarding any adverse experiences among the first 10,000 people who use a given drug. Active surveillance is conducted in Japan for the first six months of a product's use; the approach involves repeated announcements about vigilance during the early postmarketing phase and queries to doctors by trained drug-company representatives. In the United States, clinicians are exhorted, but not required, to report drug-related adverse events either to drug manufacturers or directly to the FDA; the reporting of adverse events after vaccine administration is required but still incomplete.
Other systems of active surveillance that are less dependent on the actions of individual reporters are under examination and development. These systems may examine networks of care in which drug-related adverse events occur frequently, such as liver-transplantation centers with respect to instances of severe liver injury. Bennett and colleagues have previously described how their system of networking among tertiary medical centers that offer plasmapheresis allowed them to detect cases of thrombotic thrombocytopenic purpura associated with the use of antiplatelet agents.1 Other settings of interest may include emergency departments, where instances of overdose, medication errors, and anaphylaxis may be detected. Finally, increasing the electronic capture of health care encounters, coupled with efforts to develop a common health information infrastructure throughout the United States, may allow data mining or adverse-event–detection algorithms to be applied to health systems data from both inpatient and outpatient settings. The extension of prescription-drug coverage to Medicare beneficiaries may offer the opportunity to link drug-exposure data with patterns of hospitalization or specific diseases among more than 40 million elderly persons who are at high risk for medication-related events.
Whether active surveillance or data-mining methods are used, finding potential drug-safety problems requires skillful observation by clinicians who are attuned to the possibility of drug-related adverse events and aware of the need to report them. As Bennett et al. can attest, the collective information obtained from "prepared minds" can identify modifiable risk factors and thereby minimize risk.
Source Information
From the Office of Drug Safety, Center for Drug Evaluation and Research, Food and Drug Administration, Rockville, Md.
References
Bennett CL, Connors JM, Carwile JM, et al. Thrombotic thrombocytopenic purpura associated with clopidogrel. N Engl J Med 2000;342:1773-1777.(Anne Trontell, M.D., M.P.)
In this issue of the Journal, Bennett and colleagues (pages 1403–1408) demonstrate the power and importance of the detection, analysis, and remediation of adverse drug events. When there was a sudden increase in the rate of sporadic reports of pure red-cell aplasia in association with the use of epoetin, it prompted an analysis of the worldwide adverse-event experience of pure red-cell aplasia in patients treated with epoetin. Careful analysis of the multinational clinical experience of many patients, in turn, suggested that subcutaneous administration and variable storage conditions are risk factors for epoetin-associated pure red-cell aplasia. Interventions designed to minimize these risk factors were followed by an 83 percent decrease in the number of reported cases.
Underlying the successful elucidation of this association was the recognition of a potential link between the drug and the adverse event of pure red-cell aplasia. In this specific instance and in general, recognition is aided by the degree of unexpectedness of the event, given the circumstances of the individual patient, the underlying disease, and background rates of the particular type of event. Highly unusual or infrequent outcomes (such as pure red-cell aplasia) are strong triggers of suspicion about the possible contributing role of a drug. Similarly, awareness may be heightened when an unexpected and otherwise unexplained adverse outcome is often drug-related; examples include Stevens–Johnson syndrome, bone marrow or liver failure, and torsades de pointes. If the adverse event in question occurs at a high background rate or may arise from the underlying disease, the recognition of an association with a drug is more difficult and may hinge on a finding of a strong temporal relationship between the event and drug therapy, abatement of the adverse effects when therapy is discontinued, or recurrence on reexposure to the drug. These and other factors that suggest a possible drug-related adverse event are listed in the Table.
Table. Factors Suggestive of a Possible Adverse Drug Reaction.
Once the potential linkage of an adverse event with a drug has been recognized, pertinent clinical information must be shared with a wide audience of other clinicians. The sharing and assessment of such information can occur through professional meetings, in the scientific literature, or through the reporting of adverse events to the relevant pharmaceutical company or one of its representatives. In the United States, Japan, and many other countries, including those of the European Union, drug companies must report to the appropriate regulatory authorities any adverse events that are made known to any company representative. In the United States, the public health and regulatory jurisdiction of the Food and Drug Administration (FDA) over pharmaceutical manufacturers frees clinicians from constraints with regard to the disclosure of protected health information without a patient's permission; the reporting of adverse events and postmarketing safety surveillance are explicitly exempted under the Health Insurance Portability and Accountability Act (HIPAA).
In the case of epoetin and pure red-cell aplasia, the compilation and analysis of the data on adverse events revealed risk factors that may have contributed to the development of the disease. The FDA, the World Health Organization (WHO), and other regulatory authorities maintain databases of adverse-event reports and analyze them systematically for new safety signals. The FDA's Adverse Event Reporting System, sometimes called MedWatch for the reporting form (www.fda.gov/medwatch/safety/3500.pdf) and outreach program encouraging reporting on the safety of medical products, has collected information since 1969 and now contains more than 2.5 million adverse-event reports for several thousand drug products. Within the FDA, staff members in the Office of Drug Safety review all individual case reports involving outcomes that are life-threatening, that result in death, that lead to hospitalization, disability, or congenital anomalies, or that require interventions to prevent clinically significant impairment. Targeted and comprehensive investigation and analyses may be triggered by one striking case report, by an unusual pattern of adverse events, or by a collection of adverse-event reports exceeding that expected in usual clinical experience or a given clinical context. The FDA, like the WHO and government agencies in other countries, also uses statistical methods of data mining to help discern which combinations of drugs and adverse events in their databases are distinctive. Data mining can trigger additional case-study analyses or population-based studies.
Adverse events reported by clinicians need to be placed in context so that the magnitude of the potential risk associated with a given product can be estimated. The volume of drug use often provides a crude context for the level of risk: reporting rates of adverse events may be derived using the volume of prescriptions written or dispensed as a surrogate for population exposure. Whenever possible, true population-based incidence rates are the preferred method for describing and comparing the levels of risk associated with various products; reporting rates are based only on those adverse events that are actually reported, which may represent as little as 1 percent of all cases. In general, adverse-event reporting is greatest during the first few years of a product's marketing, and reporting can be stimulated by scientific or media attention.
The ability to detect drug-safety signals through spontaneous reports is limited by the low, and highly variable, percentage of events that are reported, the uncertain estimates of exposure represented by the number of prescriptions written, and the passive nature of data collection. There are a few systems that use active ascertainment of adverse events from doctors. In the United Kingdom, reports of such events are actively solicited through the Prescription-Event Monitoring system, which surveys prescribers regarding any adverse experiences among the first 10,000 people who use a given drug. Active surveillance is conducted in Japan for the first six months of a product's use; the approach involves repeated announcements about vigilance during the early postmarketing phase and queries to doctors by trained drug-company representatives. In the United States, clinicians are exhorted, but not required, to report drug-related adverse events either to drug manufacturers or directly to the FDA; the reporting of adverse events after vaccine administration is required but still incomplete.
Other systems of active surveillance that are less dependent on the actions of individual reporters are under examination and development. These systems may examine networks of care in which drug-related adverse events occur frequently, such as liver-transplantation centers with respect to instances of severe liver injury. Bennett and colleagues have previously described how their system of networking among tertiary medical centers that offer plasmapheresis allowed them to detect cases of thrombotic thrombocytopenic purpura associated with the use of antiplatelet agents.1 Other settings of interest may include emergency departments, where instances of overdose, medication errors, and anaphylaxis may be detected. Finally, increasing the electronic capture of health care encounters, coupled with efforts to develop a common health information infrastructure throughout the United States, may allow data mining or adverse-event–detection algorithms to be applied to health systems data from both inpatient and outpatient settings. The extension of prescription-drug coverage to Medicare beneficiaries may offer the opportunity to link drug-exposure data with patterns of hospitalization or specific diseases among more than 40 million elderly persons who are at high risk for medication-related events.
Whether active surveillance or data-mining methods are used, finding potential drug-safety problems requires skillful observation by clinicians who are attuned to the possibility of drug-related adverse events and aware of the need to report them. As Bennett et al. can attest, the collective information obtained from "prepared minds" can identify modifiable risk factors and thereby minimize risk.
Source Information
From the Office of Drug Safety, Center for Drug Evaluation and Research, Food and Drug Administration, Rockville, Md.
References
Bennett CL, Connors JM, Carwile JM, et al. Thrombotic thrombocytopenic purpura associated with clopidogrel. N Engl J Med 2000;342:1773-1777.(Anne Trontell, M.D., M.P.)