Department of Health and Human Services Public Health Service Food and Drug Administration Center for Drug Evaluation and Research Office of Surveillance and Epidemiology
Date: November 19, 2010
Janet Woodcock, MD To: Director, CDER
John Jenkins, MD
Director, Office of New Drugs From: Gerald J. Dal Pan, MD, MHS
Director, Office of Surveillance and Epidemiology Subject: Updated Epidemiological Review of Propoxyphene Safety
Drug Name(s): Propoxyphene Products Application Multiple NDAs and ANDAs Type/Number:
Applicant/sponsor: Xanodyne (NDA Sponsor) and Multiple ANDA Sponsors
Note: This version corrects erroneous footnote numbering in the version signed on 18 November 2010.
1 Reference ID: 2866332
I. INTRODUCTION
This document provides a summary of a review completed in January 2010 by the Office of Surveillance and Epidemiology/Division of Epidemiology (OSE/DEPI) of epidemiological data relating to the safety of propoxyphene drug products. The OSE staff involved in the review included Fatmatta Kuyateh, MD, MS, Medical Officer; Catherine Dormitzer, PhD, MPH, Epidemiologist; Sigal Kaplan, PhD, Team Leader; Cynthia Kornegay, PhD, Team Leader; and Solomon Iyasu, MD, MPH, Director. The review was conducted as part of FDA’s ongoing safety surveillance of propoxyphene drug products following the July 7, 2009 denial of a Citizen Petition filed by Public Citizen in 2006 (the 2006 Petition), which asked the agency to withdraw propoxyphene drug products from the U.S. market on grounds that their risks outweigh their relatively weak benefits as an analgesic.
Based on a comprehensive review of the available evidence, including an advisory committee meeting held in January 2009 (2009 AC), FDA concluded that there was insufficient evidence to support withdrawing propoxyphene from the market at that time, and the 2006 Petition was denied.
Following the denial of the 2006 Petition, OSE was asked to review whether certain data sources relied on in the Petition could be re-analyzed, supplemented, and/or expanded to provide further certainty on this point. OSE also was asked to review information in several articles published both before and after FDA’s denial of the 2006 Petition. In January 2010, the OSE reviewers determined that the updated epidemiological evidence favored removal of propoxyphene from the market. The findings of that review are described below.
After consideration by a CDER-wide working group and the Drug Safety Board, CDER management decided that, although the updated evidence and analysis strengthened concern regarding propoxyphene,the available data did not provide sufficient evidence to conclude that propoxyphene, at recommended doses, was responsible for the observed findings. Further work was undertaken, including a review of clinical pharmacology and animal toxicology data by Dr. Mark Avigan in OSE, and a final decision was deferred pending review of new clinical evidence from an ongoing study of potential cardiac risks (TQT study), which, at that time, was being conducted by the NDA sponsor, Xanodyne Pharmaceuticals Inc. FDA has now reviewed the results of Xanodyne’s preliminary pharmacokinetic study, conducted to determine appropriate dosing for the TQT study, and has concluded that the data demonstrate a clear, dose-related effect on cardiac electrophysiology. The results of the new TQT study provide evidence that propoxyphene can have an adverse cardiotoxic effect at therapeutic doses. The postmarket signals, including expanded epidemiological analyses, and the findings of Dr. Avigan’s review, are consistent with this conclusion.
1 In particular, epidemiological evidence relied on in the 2006 Petition was found to be suggestive, but not conclusive, with regard to the Petition’s assertion that propoxyphene has been associated with substantial numbers of non-intentional and intentional deaths. 1
Docket No. FDA-2006-P-0270/PDN (July 7, 2009)
2
Reference ID: 2866332
This memorandum describes the epidemiological review that was conducted prior to the receipt of the new clinical evidence from the TQT study.
After review of the data in this memorandum, a review of clinical pharmacology data and animal toxicology data conducted by Dr. Mark Avigan of OSE, and the electrocardiographic results from the above-referenced TQT study, I conclude that propoxyphene and propoxyphene-containing products should no longer be marketed.
II. DISCUSSION
This memorandum describes our re-evaluation and updated analyses of epidemiological data and information. These include:
At the 2009 Advisory Committee meeting, Dr. Sidney Wolfe presented a report titled “Drugs Identified in Deceased Persons by Florida Medical Examiners” produced by the Florida Department of Law Enforcement (FDLE). He showed a graph of propoxyphenerelated deaths ranging from 328 to 368 for the period of 2003 to 2007. He pointed out that in 2007, 85 (25%) of the 341 propoxyphene-related deaths were caused by propoxyphene. 2008 Medical Examiners Commission Drug Report: http://www.fdle.state.fl.us/Content/getdoc/a37959db85e0-42f9-b6d6-cdef532f22f8/2008DrugReport.aspx
3
- Florida Medical Examiner Data
- Data from the Drug Abuse Warning Network (DAWN), including data on emergency department visits (DAWN-ED) and reports of fatalities received from medical examiners or coroners (DAWN-ME/C
- Suicide Data from England and Wales
- 1984 Case Study of Severe Propoxyphene Self-Poisoning
- 1987 Study of Effects of Governmental Warnings
- Comparative Study of Fatality Rates for Propoxyphene vs. Other Analgesics
- Case Study of Effects of UK Propoxyphene Withdrawal on Suicide Deaths
- Additional Information Reviewed by OSE
- Florida Medical Examiner Data
The FDLE releases reports annually titled “Drugs Identified by Florida Medical Examiners”. Through toxicology reports submitted to Florida medical examiners, the FDLE identifies drug-related deaths for which the drug was determined to be the cause of death, or present in the body at the time of death but not causal. A drug is only indicated as the cause of death when, after examining all evidence and the autopsy and toxicology results, the medical examiner determines the drug played a causal role in the death. Multiple drugs can be listed as the cause of death.
2 The mortality data presented by Dr. Wolfe were notable but warranted further investigation since these numbers were not to 2Reference ID: 2866332 be put into the context of use, and could not be directly compared to other mortality associated with other opioids.
OSE/DEPI requested and received similar drug-related mortality data from the FDLE (Table 1). Because the numbers reported at the Advisory Committee meeting were not placed in the context of drug exposure or availability, we could not compare the numbers of deaths to those of other opioids. As a result, OSE/DEPI conducted a more comprehensive analysis of the FDLE data. Information on total prescriptions dispensed was obtained for use as a measure of estimated exposure in Florida. The frequency of deaths was then adjusted for potential exposure by using the number of prescriptions dispensed as the denominator. Two comparator analgesics, tramadol and hydrocodone, were analyzed using the same methods to examine if similar trends of drug-related deaths were found. Tramadol was selected because it has a similar indication to propoxyphene and is currently not a scheduled opioid, so it has a similar level of control to propoxyphene, which is a C-IV level of scheduling. Hydrocodone combination products are scheduled as a C-III, and consequently expected to have more toxicity and higher levels of drug abuse than propoxyphene. Hydrocodone therefore served as a positive control. Hydrocodone also has large sales volume resulting in estimates that are more precise. In addition, tramadol and hydrocodone have been considered possible alternatives to propoxyphene.
Table 1 shows, for each drug, the total number of deaths for which the drug was present at time of death, and the number of deaths for which the drug was determined by the medical examiner to be the cause of the death. The number in parentheses is the percentage of the total deaths for which the drug was determined by the medical examiner to be the cause of death.
There were 1,740 deaths reported by the Florida Medical Examiners Office where propoxyphene products was either present or causal, 670 deaths for tramadol and 2,963 deaths for hydrocodone products for the years 2003 through 2007. Over this five-year period, propoxyphene was determined to be causal in 26% of the propoxyphene-related deaths. Tramadol was causal in 21% of tramadol-related deaths and hydrocodone in 34% of hydrocodone-related deaths. From 2003 – 2007, there were approximately 11 million prescriptions for propoxyphene, 34 million for hydrocodone and 6.6 million for tramadol. The number of prescriptions dispensed serves as a proxy for the amount of drugs available (or exposure) in the community. Over the five-year period, the number of drug-related deaths adjusted for drug utilization was approximately 16 deaths per 100,000 prescriptions for propoxyphene, 10 deaths per 100,000 prescriptions for tramadol and 9 deaths per 100,000 prescriptions for hydrocodone, the positive control. The number of deaths per 100,000 prescriptions was consistently higher for propoxyphene than the comparator drugs through all five years of data particularly when considered as causal. The two-fold difference in number of drug-caused deaths per 100,000 prescriptions for propoxyphene, compared to tramadol and hydrocodone is noteworthy. The increased ratio of deaths per 100,000 prescriptions associated with propoxyphene strongly suggests greater toxicity when compared to tramadol and hydrocodone. These results provide evidence to support withdrawing propoxyphene from the market.
4
Reference ID: 2866332
| Table 1: FDLE: Number of Drug Related Deaths Reported by Florida Medical Examiners, and Dispensed Prescriptions for Propoxyphene, Tramadol and Hydrocodone 2003 -2007, FloridaPropoxyphene |
Tramadol |
Hydrocodone |
| Number of drug-related deaths
|
1,740
|
670
|
2,963
|
| Present
a
|
1280
|
527
|
1964
|
| Causal
b
|
460 (26%)
|
143 (21%)
|
910 (33%)
|
| Number of Prescriptions Dispensed in Florida |
10,931,158
|
6,608,176
|
33,987,828
|
| Number of Total Deaths
c per 100,000 prescriptions
|
15.9
|
10.1
|
8.7
|
| Number of Causally-Related Deaths per 100,000 prescriptions |
4.2
|
2.2
|
2.9
|
Tysabri is a drug prescribed to help patients with multiple sclerosis or Crohn’s disease. Recent reports to the FDA has shown that Tysabri infusions increase the risk of progressive multifocal leukoencephalopathy (PML), an opportunistic viral infection of the brain that usually leads to death or severe disability. Cases of PML have been reported to the FDA in patients taking Tysabri who were recently or concomitantly treated with immunomodulators or immunosuppressants, as well as in patients receiving Tysabri as monotherapy.
