Safety of New Antimicrobials Approved by the FDA in the Last 7 Years
EMBARGOED UNTIL: Friday, September 13, 8:30 AM MDT
(Session 206, Paper K-1544)
Tina Khadem
State Univ. of New York at Buffalo, Buffalo, NY, United States
Email: tinakhad@buffalo.edu
Phone: 412-215-6793
In the past 7 years the U.S. FDA has approved 11 antimicrobial new molecular entities including: anidulafungin, darunavir, maraviroc, raltegravir, doripenem, telavancin, ceftaroline, boceprevir, telaprevir, fidaxomicin and bedaquiline. Historically, about 3% of FDA approved antimicrobials are removed from the US market due to untoward safety concerns. Our goal was to qualitatively and quantitatively review the FDA Adverse Event Reporting System (AERS) database to provide clinicians with a general understanding of the comparative occurrence of clinically important adverse events seen in newly approved antimicrobials approved by the FDA in the past 7 years. Bedaquiline, the most recently approved antimicrobial, was not represented due to the 5-month delay in the release of the public database. Adverse event signals were seen for 6 of the remaining 10 antimicrobial agents, many of which are not included in package inserts. They include: doripenem-associated hepatic dysfunction and hyperchloremia; boceprevir-associated weight loss; darunavir-associated premature labor, sudden infant death syndrome, ventricular hypertrophy, acute coronary syndromes, and congenital anomaly in offspring; raltegravir-associated congenital heart valve disorders, and sudden infant death syndrome. Adverse event signals identified but already listed in the package insert include: telavancin-associated acute renal failure; doripenem-associated seizures and thrombocytopenia; boceprevir-associated anemia; maraviroc-associated malignancies and myocardial infarction; and raltegravir-associated hepatitis. There appears to be adverse events associated with these antimicrobials approved in the last 7 years, including those not identified in package inserts. However, investigative studies are needed to further explore these statistical associations.
In this retrospective review of the FDA AERS database, Empirica Signal software was used to query all adverse event-drug combinations reported from November 1968 to December 31 2012. We calculated a disproportionality statistic, namely the Empirical Bayesian Geometric Mean (EBGM), for reports of antimicrobial adverse events. With use of a disproportionality analysis such as the EBGM, each antimicrobial agent was compared with all agents within their antimicrobial therapeutic class. The FDA considers an EBGM significant if the 5th percentile of the distribution is at least two (EB05 > 2.0). This review was conducted by the previously specified authors including Tina Khadem (presenting author), Robbert Van Manen, and Jack Brown (senior author) and will be presented at the Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) in Denver, CO on September 13, 2013. There was no external funding for this study.
Although clinical trials serve as the gold standard for demonstrating the safety and efficacy of newly approved drugs, unfortunately they allow for study of only a homogenous patient population for a finite period of time and likely do not reflect real-world use. Furthermore, the number of patients enrolled in drug development trials is not adequate to determine rare, yet serious, adverse events. Databases for spontaneous reporting of adverse reactions such as the FDA AERS are useful mechanisms for gaining insight into such rare adverse events. However, a limitation of AERS is the lack of a quantitative exposure estimate to put such rare events into perspective. One tool that may help overcome this limitation is disproportionality analysis. However of note, disproportionality analysis of a spontaneous report does not generate an incidence that may be generalized to a population but rather a signal that quantifies a potential correlation. Therefore, a direct cause and effect relationship cannot be inferred and further studies are warranted to investigate the potential correlation.