|
|
||
 |
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology Dissemination of Information on Potentially Fatal Adverse Drug Reactions for Cancer Drugs From 2000 to 2002: First Results From the Research on Adverse Drug Events and Reports Project
From the Veterans Affairs Midwest Center for Health Services and Policy Research, Lakeside Division, the Division of Hematology/Oncology of the Department of Medicine, the Center for Healthcare Studies, and the Robert H. Lurie Comprehensive Cancer Center, Northwestern University Medical Center; the Division of General Internal Medicine, Department of Medicine, The Feinberg School of Medicine, Northwestern University; the Center for Pharmacoeconomic Research, College of Pharmacy, University of Illinois at Chicago, Chicago, IL; the Veterans Affairs Salt Lake City Health Care System and the University of Utah School of Medicine, Salt Lake City, UT; the Stanley S. Scott Cancer Center and Section of Hematology/Oncology, Louisiana State University Health Sciences Center, New Orleans, LA; and the Veterans Affairs Cooperative Studies Program Clinical Research Pharmacy Coordinating Center, University of New Mexico, Albuquerque, NM. Address reprint requests to Charles L. Bennett, MD, PhD, VA Lakeside Medical Center, Division of Hematology/Oncology, Department of Medicine, 400 E Ontario St, Suite 205, Chicago, IL 60611; e-mail: cbenne{at}northwestern.edu.
Purpose: To describe the clinical findings, occurrence rates, causality evidence, and dissemination media for serious cancer drug–associated adverse drug reactions (ADRs) reported in the postmarketing setting. Methods: ADRs were termed serious if they resulted in death or severe organ failure. ADR information for oncology drugs from package insert (PI) revisions, so-called Dear Doctor letters, and journal articles was evaluated to identify serious ADRs reported from 2000 to 2002. Timing and content of information disseminated was assessed. Results: Twenty-five serious ADRs associated with 22 oncology drugs were identified after approval. Approximately half of these serious ADRs are associated with drugs approved before 1995. ADRs were described in articles in medical journals (17 ADRs), PI revisions (18 ADRs), and Dear Doctor letters (12 ADRs). PI revisions occurred less than 1 year after peer-reviewed publication for four ADRs. These revisions often differed for similar ADRs that occurred with drugs of the same class. Five of the seven ADRs lacking PI changes occurred with off-label use, for which PI change is not recommended by US Food and Drug Administration (FDA) policy. No cancer drug was withdrawn from the market during the observation period. Conclusion: Our findings demonstrate that serious ADRs may be discovered as long as 36 years after a drug receives FDA approval. This suggests a need for continued vigilance and efficient strategies for dissemination of information about ADRs associated with cancer drugs.
ADVERSE DRUG reactions (ADRs) related to cancer drugs are an important cause of morbidity and mortality. Serious or potentially fatal ADRs are often detected after cancer drugs are widely used in oncology practice.1 Premarketing clinical trials are designed primarily to identify benefits and common side effects of new drugs; however, the size of these studies generally does not exceed 3,000 patients, limiting the likelihood of detecting rare ADRs before approval.2 Oncology drug–associated ADRs are particularly important to consider, given that these drugs may be especially likely to cause ADRs because they are designed to be cytotoxic and thus often injure normal cells in addition to the malignant cells. When previously unidentified but serious ADRs are reported after the drug has been approved by the US Food and Drug Administration (FDA), information dissemination occurs through revised package inserts (PIs), so-called Dear Doctor letters, and/or publications in medical journals. Despite medical professionals’ and patients’ dependence on this information to ensure safe pharmaceutical usage, ADR reporting is often delayed and inconsistent in format.3 We report identification and dissemination media used to describe 25 recently reported serious ADRs associated with 22 commonly used cancer drugs.
We reviewed information for all serious ADRs associated with oncology drugs reported from 2000 to 2002, using the following search strategy. An ADR was characterized according to the WHO definition as "a response to a drug that is noxious and unintended and occurs at doses normally used in man for the prophylaxis, diagnosis or therapy of disease, or for modification of physiological function."4 According to the FDA and the WHO, a serious ADR is "any untoward medical occurrence that at any dose results in death, requires hospital admission or prolongation of existing hospital stay, results in persistent or significant disability/incapacity, or is life threatening."4 However, in this article, we restricted our analyses to those serious ADRs that were potentially life threatening and excluded serious ADRs that prolong hospitalization without a life-threatening component.4,5 Operationally, these included ADRs for which: (1) 10 or more of the initially reported cases of the ADR resulted in death; required rapid medical intervention, such as intubation, dialysis, plasmapheresis, or cardiopulmonary resuscitation; or required emergency hospitalization for treatment; or (2) the toxicity was believed to be potentially life threatening (by unanimous consensus of all five oncologist coauthors of this study) and (A) the FDA and the pharmaceutical company negotiated descriptive language included in a Dear Doctor letter mailed by the pharmaceutical supplier, or (B) the toxicity was described in revisions in sections of the PI generally associated with the most serious drug-associated adverse events (ie, the so-called black box warnings, warnings sections, or contraindications sections). Serious ADRs that fulfill these criteria were identified through an exhaustive search of all PI changes and Dear Doctor letters from 2000 to 2002 on the FDA’s MedWatch Web site (20 ADRs were identified). Serious ADRs identified during this time period within the investigative team’s comprehensive cancer center that were not identified from PI changes or Dear Doctor letters were also included (five ADRs). Additional information about identified ADRs was obtained from reports from Medline, pharmacovigilance programs of pharmaceutical manufacturers of cancer drugs, the FDA’s MedWatch program, and Dear Doctor letters sent by pharmaceutical manufacturers to physicians in the United States.6 MedWatch reports, which are submitted voluntarily by health professionals and on a mandatory basis by pharmaceutical manufacturers to a centralized database at the FDA, were obtained through the FDA’s Adverse Event Reporting System.
For each drug, information was obtained on the original FDA-approved indication, the number of reported serious and fatal instances of the specific drug-associated ADR, clinical findings, evidence supporting causality, and estimates of the frequency. When the ADR was reported in a clinical trial, rates were estimated from these data (as a ratio of the number of cases to the number of patients receiving the drug in clinical trial; eight ADRs). Other rate estimates were derived from the literature (four ADRs) or pharmaceutical supplier estimates (one ADR). When sufficient data were not available to calculate a rate estimate, consensus opinion of the five oncologist coauthors was used to assess whether the rate was likely to be common ( Data related to ADR information dissemination in Dear Doctor letters, product label revisions, and peer-reviewed articles were reviewed. For each ADR, dates were identified for the following events: drug approval (to approximate the date of initial marketing), publication of a peer-reviewed article describing more than 10 cases of suspected drug-associated ADR, pharmaceutical manufacturers’ PI revisions, and Dear Doctor mailings. The time (to the nearest month) that elapsed between FDA drug approval and ADR reporting in a PI revision, Dear Doctor letter, or published case series was measured.
We identified 25 serious ADRs, 12 occurring with drugs approved before 1995, six with drugs approved via standard approval after 1995, and seven with drugs approved via accelerated approval. These ADRs were associated with 22 oncology drugs, five of which received accelerated FDA approval. The most common toxicities were severe infusion reactions (n = 3 [13%]) and interstitial pneumonitis (n = 3 [13%]). ADRs were identified within 2 years of FDA approval (nine ADRs), more than 2 years but less than 6 years after FDA approval (six ADRs), and between 6 and 36 years after FDA approval (10 ADRs; Table 1  ). ADRs varied in times of onset, clinical characteristics, and estimated rates. Adverse events occurred immediately or shortly after initiation of drug therapy (five ADRs), after 1 day to 1 week of drug therapy (two ADRs), after 1 week to 1 month of drug therapy (three ADRs), after 1 to 3 months (10 ADRs), after more than 3 months of use (five ADRs), and after years of use (one ADR; Table 1, Fig 1). Deaths from an individual ADR were reported for as many as 44 individuals (for irinotecan).
Factors supporting causality were onset during or shortly after completion of infusion (seven ADRs), occurrence of a toxic reaction identified previously with a chemically related drug (10 ADRs), identification of antibodies against the drug (one ADR), measurement of elevated levels of tumor necrosis factor and interleukin-6 while an infusion reaction occurred (one ADR), case-control epidemiologic studies (two ADRs), significantly higher rates in the treatment versus control arms of phase III clinical trials (three ADRs), greater than 10% incidence rate in phase I or II clinical trials (three ADRs), and large numbers of cases identified in reports from centers that treated large numbers of patients with a specific drug (five ADRs). (Each supporting document is referenced in Table 1  .)
Rate estimates for relatively common, serious ADRs were derived from phase I or II clinical trials (four ADRs), phase III clinical trials (two ADRs), and data from experienced centers (two ADRs), and ranged from one in 30 to one in three. For infrequent ADRs, rates were usually obtained from the pharmaceutical manufacturer (rituximab-associated infusion reactions) or from the literature (four ADRs),7–9 ranging from one in 170 to one in 13,300 (Table 1
Revised PIs, Dear Doctor letters, and peer-reviewed published articles described the ADRs included in this report (Table 1 Nineteen serious ADRs were described in 21 PI revisions, with nine as black box warnings, eight as warnings (Procrit [epoetin alfa; package insert], Bridgewater, NJ: Ortho Biotech LP, 2002) or precautions, two as adverse reactions, one as a contraindication, and one as a dosage and administration recommendation.10–27 PI revisions occurred within 2 years after FDA approval (six ADRs), more than 2 years but less than 6 years after FDA approval (six ADRs), and between 6 and 36 years after FDA approval (nine ADRs). Drugs with serious ADRs not described in the PI include one associated with an off-label administration schedule (overly rapid infusion of intravenous immune globulin), one with off-label splitting of single-use vials among multiple patients (erythropoietin), two with use in off-label clinical settings (thalidomide for cancer patients and oprelvekin for pediatric cancer patients), and one rare ADR associated with on-label use (flutamide-associated interstitial pneumonitis). For warfarin interaction with capecitabine, two PI changes were made, with the first change in the precautions section and the second in a black box.21 Two changes were also made for pamidronate-associated severe renal toxicity; the first in the dosage and administration and the second in the warnings section.16,27 Peer-reviewed journal articles describing case series of 17 serious ADRs were published within 2 years of FDA approval (seven ADRs), between 2 and 6 years after FDA approval, (four ADRs), and between 6 and 23 years after FDA approval (six ADRs).7–9,28–40 Six of the 11 ADRs (60%) described in black box warnings or contraindications were described in peer-reviewed articles.9,32–34,36,38 Two ADRs (thalidomide-associated thrombophlebitis and flutamide-associated pneumonitis) were described in journal articles, but not in Dear Doctor letters or in revised PIs.7,8,39 Six ADRs were described in journal articles 1 year or more before PI revisions occurred.30,32–35,38 In the United States, pharmaceutical manufacturers and distributors mailed Dear Doctor letters to physicians describing 12 serious ADRs, which occurred within 2 years of FDA approval (four ADRs), more than 2 years and less than 6 years after FDA approval (three ADRs), and between 6 and 25 years after FDA approval (five ADRs).41–51 Seven of the 11 ADRs (64%) described in black box warnings or contraindications are described in Dear Doctor letters. One ADR, erythropoietin-associated pure red cell aplasia, was described in Dear Doctor letters and PI revisions in Europe, Canada, and Australia in 2001, but not until 2002 in the United States. One ADR, oprelvekin-associated papilledema in children, was described in a Dear Doctor letter, but not in the PI nor in a peer-reviewed journal article.45
More than half of all pharmaceutical agents have serious ADRs identified after they have received FDA approval for marketing.52 In this study, we found that 25 serious ADRs associated with 22 oncology drugs were identified from 2000 to 2002 by pharmaceutical suppliers or investigators associated with this study. The information related to these serious ADRs was disseminated to the medical community through a variety of sources, primarily Dear Doctor letters, revisions in the FDA-approved PIs, and publications in the medical literature. The dissemination efforts were variable with respect to content, timing, and types of media. In interpreting our findings, several factors should be considered. Sources of the case series information were related to ADR frequency. Case series information for common ADRs was obtained primarily from clinical trial reports. MedWatch, the FDA’s voluntary reporting system, provided the majority of information for the rare ADRs. Only between 1% and 10% of all serious ADRs are estimated to be reported to MedWatch.53 Five ADRs were reported from cancer center investigators who treated large numbers of patients with the relevant drug. Erythropoietin-associated pure red cell aplasia was reported in Europe, Canada, and Australia in association with a first-generation erythropoietin product manufactured in Puerto Rico, but not until 2002 with the first- or second-generation erythropoietin products manufactured in the United States. Tamoxifen-associated endometrial adenocarcinomas were initially reported in a case-control study from Belgium.33 In several European countries, because of concern about low rates of voluntary reporting, enhanced surveillance systems have been developed.54–56 In some cases, clinical pharmacology considerations predicted that a serious ADR should be expected.57 Marked elevations in prothrombin times were noted with concomitant use of warfarin and the fluorouracil prodrug capecitabine, a finding similar to that reported with warfarin and fluorouracil.58 Interstitial pneumonitis was reported with nonsteroidal antiandrogens and vinca alkaloids.7,35 With the emergence of biologic therapies, new types of class effects have been identified. Pure red cell aplasia (reported in chronic renal failure patients who received recombinant erythropoietin in Europe, Australia, and Canada) and severe thrombocytopenia (reported in volunteers who received recombinant human megakaryocyte growth and development factor) are caused by neutralizing antibodies.35,59,60 Another toxic class effect, infusion reactions (which was identified with the monoclonal antibody products gemtuzumab, rituximab, and trastuzumab), appears to be secondary to cytokines released as antibodies bind to circulating antigen-expressing cells and are then removed in the reticuloendothelial system of the lungs, spleen, and liver.34 Serious ADRs emerged throughout the life cycle of oncology drugs. Imatinib-associated hemorrhage, presumably related to tumor degeneration, is described in a PI revision made 8 months after imatinib’s initial accelerated FDA approval.37 The ADR information was added when an additional FDA approval was obtained for the treatment of gastrointestinal stromal tumors. It is not surprising that this serious ADR was not identified earlier, because at the time of the initial FDA approval of imatinib (for chronic myelogenous leukemia), clinical trials had just begun for persons with gastrointestinal stromal tumors. In contrast, severe neuropsychiatric disorders with interferon alfa-2a, endometrial adenocarcinoma and uterine sarcoma with tamoxifen, warning against intrathecal administration for vinblastine, and severe bone marrow suppression for thioguanine were described in PI revisions between 15 and 36 years after the first FDA approval. Similar experiences of delayed ADR reporting for drugs used in settings other than oncology include marked hypotension with clozapine, pulmonary fibrosis with tocainide, subarachnoid hemorrhage with phenylpropanolamine, and thrombotic thrombocytopenic purpura with ticlopidine.37,61,62 Inconsistencies in ADR reporting, particularly in PI changes, were noted in our study. One fourth of the serious ADRs in this study are not described in PIs; most of the ADRs not included in PIs occur with off-label use, for which PI change is not FDA policy. In the United States, pure red cell aplasia was described in the PI in 200163 for darbepoetin, a second-generation form of erythropoietin, but not until 2002 was it added to the PI for first-generation erythropoietin products sold in the United States, although few cases of pure red cell aplasia have been reported for either of these products.64 In addition, inconsistencies in the ADR information in the PI were apparent for nonsteroidal antiandrogen–associated interstitial pneumonitis (described in the PI for bicalutamide, but not for flutamide) and vinca alkaloid–associated pneumonitis (described as a warning for vinorelbine but not in the PIs for vincristine and vinblastine).
Our findings have implications related to accelerated approval of cancer drugs.57,65 Accelerated approval is one of three FDA processes established under the Prescription Drug User Fee Act of 1992 and the Food and Drug Administration Modernization Act of 1997 designed to improve access to new cancer drugs. (The others are fast-track designation and priority review66; Table 2
Our findings have several policy implications. First, ADR identification and reporting could be improved. Postmarketing surveillance focusing on safety should be strongly considered for cancer drugs that receive accelerated FDA approval. Targeted surveillance efforts should be considered for a new drug in the same chemical class as an older drug when there is a possibility of a class effect. To facilitate postmarketing surveillance, specific markings should be added to the PIs of drugs that receive accelerated FDA approval. Second, current ADR data collection techniques could be improved to encourage more frequent and higher quality reporting. MedWatch forms should be modified to facilitate improvements in the quality of reports, decrease the likelihood of obtaining incomplete reports, decrease the frequency of supplemental requests, and allow for better follow-up. MedWatch reports should contain guidelines to provide sufficient information to establish the causal link between the drug and the adverse event. For previously described ADRs, the FDA should post disease-specific criteria on its Web site when possible to facilitate credible establishment of a diagnosis by physicians who have sufficient information. Third, information dissemination could be improved through increased sharing of ADR information between agencies, pharmaceutical companies, and physicians. Empirical data on how often physicians obtain ADR information for cancer drugs from PIs are not available72; however, it has been suggested that physicians do not always adhere to PI or Dear Doctor letter recommendations. For the nonsteroidal analgesic bromfenac, it was noted that physicians frequently did not adhere to recommendations against extended use outlined in a revised black box warning, and the drug subsequently was withdrawn from the market.57,62 In contrast, high levels of compliance have been noted with a black box warning related to drug interactions with the antibiotic moxifloxacin.73 Information dissemination could be improved by mailing Dear Doctor letters when black box warnings or contraindications are added to PIs or releasing Internet articles ahead of print for articles that describe serious ADRs on a global scale; procedures should be developed for international sharing of ADR information. Finally, our findings do not indicate that accelerated approval has resulted in premature release of unsafe cancer treatments; rather, the results highlight the importance of timely completion of postmarketing studies to better assess the safety and efficacy profile of drugs that receive this form of FDA approval. Some limitations of our study should be considered. First, the search strategy used to identify serious ADRs for this study included only PIs, Dear Doctor letters, and experience within our comprehensive cancer center. A comprehensive MEDLINE search for all oncology drugs could identify additional serious ADRs. Other potential information sources, such as case reports or cooperative group reports, were not used in this search process. Many of the ADRs discussed here as having no case series journal articles are described in at least one case report in the published literature. These reports may provide important information for physicians but were excluded here. Second, because physicians only report a fraction of the ADRs they encounter in clinical experience, our data set is likely not to reflect all oncology drug–related ADRs that occurred during this time period. Third, pharmaceutical manufacturers focus on ADRs that are associated with on-label use because a description of ADRs associated with off-label use is thought to imply that the off-label use is approved by the FDA. More than half of oncology drug use is for off-label indications.74 Finally, because of the relatively small number of ADRs analyzed here, it is difficult to suggest trends in ADRs associated with drugs receiving accelerated versus standard FDA approval. Additional research into this area will be important in shaping policy decisions about FDA approval processes. In conclusion, our study raises important concerns related to pharmaceutical safety. Improvements in information dissemination efforts related to serious cancer-related ADRs, especially those that occur with off-label use, are needed.
The authors indicated no potential conflicts of interest.
Presented in part at the 39th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, May 2003.
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
1% occurrence rate) or rare (< 1% occurrence rate; eight ADRs). Sensitivity analyses indicated that the findings were similar for a cut-point that ranged from 0.5% to 2.5% for the designation of common versus rare. 
