Call for full disclosure of drug data before registration
EXPERTS are calling for an overhaul of drug regulatory systems following revelations about the impact of unpublished trial data and how its disclosure can impact the efficacy and safety of registered drugs.
The concerns follow a landmark investigation into the SNRI antidepressant reboxetine (Edronax) that found the drug to be “ineffective and potentially harmful”, despite almost 10 years on the Australian market and more than half a million prescriptions.
German researchers for the first time extracted the totality of data on reboxetine – three quarters of it previously unpublished.
They found reboxetine offered no significant benefits over placebo, and was inferior to SSRIs for depression remission and response rates.
Pfizer, the drug’s manufacturer, said it would review the meta-analysis before commenting further. However, the company noted in a statement that the drug was effective in preventing the relapse of depressive symptoms, and that regulators carefully balanced the risks and benefits of drugs and reflected all important safety information in the labelling.
Professor Philip Mitchell, head of psychiatry at the University of NSW, said the evidence supporting reboxetine had “disappeared” in light of the new data, and there was no longer any clinical rationale to prescribe it.
“This is probably one of the most dramatic and colourful examples of the failure to publish negative studies. This is a problem across the whole of medicine, not just psychiatry,” he said.
Dr Nik Zeps (PhD), a member of the NHMRC’s health ethics committee, believed the problem was not confined to one drug or manufacturer.
Despite the 2007 creation of clinical trial registers, it was unclear how much obligation was on companies to produce post-marketing data for drugs that had previously been approved, he said.
With such studies being costly and potentially damaging, Dr Zeps called on governments, regulators and journals to take greater responsibility for demanding full and accurate data.
“It would be disingenuous of anyone to say they’re surprised this is how drug companies are behaving... It’s what you’d expect them to do within the framework we currently have. Perhaps it’s time for an overhaul.”
The BMJ, which published the reboxetine study, said it called into question “the entire evidence synthesis enterprise”.
The journal would devote a future issue to re-evaluating treatments using unpublished data.
“Efforts are needed to restore trust in existing evidence... It is time to demonstrate a shared commitment to set the record straight,” the editors said.
Reboxetine was licensed for marketing in Australia in 2001. Usage peaked in 2006, with about 100,000 prescriptions in the calendar year, PBS data show.BMJ, online October