Thursday, August 09, 2012


In an interesting article in the on-line edition of Thrombosis and Hemostasis, July 26th 2012, Drs Victor Serebrauny ( John Hopkins University ) and Dr Dan Atar ( Oslo University Hospital, claimed that the sponsors of 3 major ( important ) clinical trials intentionally mishandled trial data to make the sponsors drug look better. The 3 trials involved were TRITON TIMI 38, RECORD and PLATO.
TRITON TIMI 38 was the trial that compared Prasugrel ( Eli Lily ) against Clopidogrel and showed that Prasugrel was better then Clopidogrel in CV outcomes, except that there is a higher incidence of major bleed. RECORD is a trial comparing Rosiglitazone ( now withdrawn ) with Metformin + Sulphonyurea to show that Rosiglitazone was not any worse then M+S in CV outcomes. PLATO is the latest trial comparing Ticagrelol against Clopidogrel, showing Ticagrelol being better then clopidogrel for CV outcomes.
Dr Serebruany and Dr Atar found that the the number of MIs ( heart attacks ) reported by the clinical events committee of the 3 trials differed from the total numbers reported from the individual sites, always favouring the sponsors drug.

Central adjudicating committees applied three strategies in MI reporting, which improved the MI outcomes with the investigational drugs, they report.
  • In TRITON TIMI-38, the MI definition was switched to a more liberal one toward the trial end, and the "extra" MIs were distributed equally to the prasugrel and clopidogrel (comparator) arms, making the numeric differences significant (from 72 to 144 MIs).
  • In PLATO, the CEC found an extra 45 MIs in the clopidogrel (comparator) arm but none in the ticagrelor arm.
  • In RECORD, 16 "silent" MIs were disallowed.
"Without this adjudication . . . the three trials' primary efficacy outcomes were not significant," Serebruany told heartwire.

This is very disturbing as it encroaches on the very principle of honesty and integrity, that all  trialist must adhere to. It I cannot trust you reporting the numbers, how can I trust your results?

I wonder how often these things happen. This one was picked up by Dr Serebruany and Atar. I wonder how many are missed.

Of course, we also give the trial investigators their right of reply.

 "There is no 'discrepancy,' " Dr Christopher Cannon (Harvard Medical School, Boston, MA), senior investigator of the TIMI study group, told heartwire. "There were many periprocedural MIs that are only diagnosed via review of the centrally run biomarkers post-PCI, and thus the investigators are not supposed to find these MIs; only the CEC is," he noted. "Thus, it is a simple explanation and totally different from the wild and uninformed speculation of the authors of this very biased piece."
 "Event adjudication is an important aspect of trials, and we do need to have open dialogue about the issues raised by the authors," Dr Kenneth W Mahaffey (Duke Clinical Research Institute, Durham, NC), CEC chair for PLATO, concurred to heartwire. However, the "current analyses were flawed, because the authors did not use the patient-level data," he noted. "In PLATO, there were events that were reported by the investigators that were not called 'events' by the CEC, because they did not meet the protocol definition for the end point, [and in turn] the CEC found events that were not reported by the investigators as end points."

At my level, I am not sure who is right and who is wrong. What I take away, is that double blind control trial, is a good idea, but it is not really blind. After all the events and outcomes have been recorded, there is a central committee who then decides which events "fits" the definition. This will allow for "massaging" of results. It is NOT so blind after all.

Real "Evidence based medicine", is there such a thing?

Lately we have also heard ( in the mass media ), that Pfizer have been paying brides to get their product to the out of US market. Now, this is very disturbing as well. The pharmas are under close scrutiny. 

It is really confusing. What and who are we to belief now?

No comments: