In deciding that the pain drug propoxyphene should come off the market, Center for Drug Evaluation and Research officials went from believing there was insufficient evidence of cardiac harm at therapeutic doses to concluding there was no way the risks of heart arrhythmia could be managed and mitigated.

The agency’s decision to opt for an immediate, rather than a phased, withdrawal of the opioid was emblematic of CDER officials belief that the risk/benefit profile for the 50-year old product had shifted so dramatically that no risk management program would be adequate.

“It is not possible to monitor for, or mitigate, the risk of fatal cardiac arrhythmia that may occur within the recommended dosing range for propoxyphene,” Division of Anesthesia and Analgesia Products Deputy Director Sharon Hertz and Office of Surveillance and Epidemiology Associate Director and Division of Pharmacovigilance I Acting Director Mark Avigan say in a joint Nov. 18 memo.

NDA holder Xanodyne Pharmaceuticals is withdrawing its branded versions of the opioid, marketed under the Darvon and Darvocet labels, at FDA’s request. The agency also requested generic manufacturers voluntarily remove their products from the market. There are currently six single-entity propoxyphene products and 22 propoxyphene/acetaminophen combination products in the U.S., according to FDA.

First Withdrawal Based On FDAAA Study

CDER leadership reached its withdrawal decision based on results of a QT study required under the FDA Amendments Act. Consequently, propoxyphene, which was first approved by FDA in 1957, has the ignominious distinction of being the first drug withdrawn on the basis of a FDAAA-required post-marketing safety study (Also see "Propoxyphene Marketing Halted By FDAAA-Required QT Safety Study" - Pink Sheet, 19 November, 2010.).

The withdrawal caps a three-decade battle by Public Citizen’s Health Research Group to pull the drug off the market. The group has argued propoxyphene is only minimally effective and poses safety risks in the form of addiction and cardiotoxicity, resulting in thousands of deaths.

FDA’s withdrawal also follows similar actions initiated by U.K. and European regulators in 2005 and 2009, respectively.

As with Avandia, CDER leadership was clearly uncomfortable relying on epidemiological and other nonclinical data to make a withdrawal decision.

How FDA reached its decision in some ways mirrors its more publicly scrutinized regulatory action on another drug, GlaxoSmithKline PLC’s thiazolidinedione Avandia. In both cases, the agency was urged to withdraw the drug by outside parties; it brought the issues before an advisory committee and its internal Drug Safety Oversight Board, and OSE reviewers believed the epidemiological evidence supported a finding of increased harm relative to comparator agents, thereby justifying withdrawal.

With Avandia, FDA allowed the diabetes drug to remain on the market under restricted distribution measures, and it ordered an independent re-adjudication of cardiovascular events in the open-label RECORD trial (Also see "FDA, EMA Decisions On Avandia Reflect The Power Of REMS" - Pink Sheet, 27 September, 2010.).

Though the final regulatory actions on Avandia and propoxyphene differed, the two examples are analogous in that FDA officials clearly were not comfortable making a withdrawal decision based on epidemiological and other nonclinical data.

While agency leaders await what they hope will be a more definitive clinical evaluation of Avandia’s CV effects through the RECORD re-adjudication, the propoxyphene QT safety study provided the “missing link” of evidence on the opioid’s cardiotoxic effects at therapeutic doses that ultimately spurred FDA’s withdrawal decision (see chart).

Continued Availability Reconsidered

FDA’s most recent examination of propoxyphene’s risk/benefit profile resulted from Public Citizen’s 2006 request seeking a phased withdrawal of the drug. That petition led to a January 2009 review by the Anesthetic and Life Support Drugs and Drug Safety and Risk Management Advisory Committees, which voted 14-12 against continued marketing (Also see "Public Citizen Takes On Propoxyphene Again, This Time Comes Out On Top" - Pink Sheet, 9 February, 2009.).

However, in July 2009 FDA denied Public Citizen’s petition, saying that although it agreed there was evidence of adverse cardiac effects with overdose levels, data were lacking that the drug was cardiotoxic at therapeutic doses.

FDA invoked its powers under FDAAA to strengthen warnings on overdose risk and require a Risk Evaluation and Mitigation Strategy with a Medication Guide. The agency also required a thorough QT (tQT) study and initiated plans for new epidemiological reviews using Medicare, Medicaid and HMO databases (Also see "A Lack Of Data Keeps Propoxyphene On The Market ... For Now At Least" - Pink Sheet, 13 July, 2009.).

Public Citizen petitioned for reconsideration, leading to further OSE analysis of emergency department and postmortem data from several sources, which showed that propoxyphene was associated with a higher frequency of deaths compared to tramadol and hydrocodone. OSE reviewers concluded these data provided adequate evidence to support product withdrawal.

No Consensus At DSOB

However, the agency’s Drug Safety Oversight Board, which considered the data in January 2010, failed to reach consensus.

“Some board members expressed the belief that the data simply represented confirmation of propoxyphene’s toxicity among patients likely to be abusing and therefore overdosing on the product,” Hertz and Avigan say in their memo. “Among board members who felt the data did affect the risk-benefit assessment, it was unclear as to the weight to give the medical examiner data. The board noted that, although propoxyphene might have unique toxicities, withdrawal from the market would have a negative public health impact …. The board feared that many patients would switch to alternate therapies for pain relief (e.g. NSAIDs, hydrocodone), which might result in exposure to other drug risks.”

Among its recommendations, the DSOB urged CDER to consider how to reduce the risk of propoxyphene use in patients who abuse the drug. In light of the board’s recommendations, management of OSE and the Office of New Drugs determined that the data did not provide sufficient evidence to conclude that propoxyphene, at recommended doses, was responsible for the observed excess mortality, Hertz and Avigan say.

In an April 2010 memo, Avigan outlined three regulatory options: deferral of a regulatory decision pending completion of new epidemiological studies, expected to take at least one to two years; deferral pending completion of a battery of studies characterizing propoxyphene’s cardiotoxic profile compared to other opioids, in addition to the tQT study in healthy volunteers required under FDAAA; or initiation of a phased withdrawal without waiting for completion of new studies.

OND and OSE senior management decided to defer any withdrawal decision until additional information from the tQT study was available.

Multiple Ascending Dose Study Spells Doom

In fact, it was a multiple ascending dose (MAD) study, conducted in healthy volunteers to determine a safe dose for the tQT study, that ultimately doomed propoxyphene.

Significant QTc interval prolongations were observed in cohorts of subjects receiving 600 mg a day (the present maximum labeled daily dose) and 900 mg. There also were dose-dependent prolongations of the PR and QRS intervals, indicating significant sodium channel blockade.

The MAD study data proved to be the tipping point for FDA’s risk/benefit assessment.

“The analgesic benefits and potential safety risks of propoxyphene products have long appeared to be very closely balanced,” Hertz and Avigan write. “However, the results of the MAD study, required under FDAAA, have added substantially to the knowledge of the effects of propoxyphene and norpropoxyphene on human cardiac electrophysiology. The findings of prolongation of the PR interval and the QT interval and widening of the QRS complex, in a dose-related manner, were present at doses within the therapeutic dosing range. This provides a previously missing link between suggestive, but not persuasive or conclusive, evidence from earlier reviews.”

Clinical data suggest even small changes in metabolic status, concomitant drug use or renal function could increase the risk of propoxyphene’s cardiotoxicity.

The clinical data suggest that even small changes in a patient’s metabolic status, concomitant drug use or renal function could increase the risk of cardiotoxicity, the FDA officials say.

Withdrawing Now, Not Later

The product withdrawal, which is proceeding immediately rather than in a phased fashion, means current propoxyphene users will have to be channeled to other analgesic products, which pose their own safety risks.

In an e-mailed response to questions from “The Pink Sheet,” FDA said the nature of the MAD study findings “was of sufficient concern that a phased withdrawal over many months was not considered an appropriate action. The cardiac electrophysiological findings of PR prolongation, QRS widening and QT prolongation that occurred with dosing at the upper end of the labeled dose range and at 50% above this dose were sufficiently large to warrant the action we chose.”

By Sue Sutter

Cardiac Risks Of Propoxyphene Raced Out Of Control As FDA Moved For Immediate Withdrawal

More from Archive

More from Pink Sheet