Disagreements Between Clinical Trial Oversight Committees and Physicians can Affect Drug Trial Results
ATLANTA, GA -– After reviewing the data from four of the
largest multi-center randomized clinical trials of new heart
disease drugs, Duke University Medical Center researchers have
found significant differences in heart attack rates reported by
physicians who actually provide the patient care and the
physician committees responsible for later determining whether
a heart attack actually occurred.
The Duke researchers said these differences, which occurred
in up to one-fifth of the cases, increased the perceived
benefits of the drug being tested in one trial and decreased
the perceived benefits of the drug in the other three trials.
While these differences of opinion had no effect on patient
care, the researchers believe that the results should help them
refine the way future trials are designed and carried out.
Most large clinical trials use clinical events committees
(CEC) to review case records to ensure that all the physicians
at participating hospitals follow the same guidelines. CECs are
usually composed of cardiology fellows and attending
cardiologists and their rulings are final.
Differences of opinion have become common as large
multi-center trials use more subtle measurements, or endpoints,
to determine the effectiveness of a particular treatment.
"It is very important for us involved in clinical trial
design to understand why such significant differences between
the physicians and the committees exist, and to make changes,"
said Duke cardiologist Dr. Kenneth Mahaffey. "Because
physicians treat their patients differently from hospital to
hospital, and from country to country in clinical trials, we
need to have a systematic way of reporting heart attacks in
In the 1980s, death was a common measurement for clinical
trials, an endpoint that leaves no room for disagreement among
physicians, the researchers said.
"However, in the 1990s, we started using such endpoints as
recurrent infarction (heart attack) to determine whether or not
a drug was effective," Mahaffey said. "While most cases of
reinfarction are obvious, in up to about 20 percent of the
cases, there were differences of opinion whether or not
reinfarction actually occurred in the clinical trials, which is
similar to what physicians see every day in practice."
Mahaffey prepared the results of the Duke study for
presentation Tuesday at the 47th annual scientific session of
the American College of Cardiology.
The four trials reviewed by the Duke researchers were called
EPIC, IMPACT-II, GUSTO-IIb and PURSUIT. Researchers from the
Cleveland Clinic coordinated the EPIC trial, the other three
were coordinated by the Duke Clinical Research Institute, of
which Mahaffey is a member.
In the EPIC trial, for example, physicians in the field
reported that 9 percent of their patients has suffered a
reinfarction, but upon further review by the CEC, the rate
turned out to be 8.3 percent. For the other three trials, the
effect was reversed:
IMPACT-II: Physicians reported 5.5 percent rate; the CEC 9.2
GUSTO-IIb: Physicians reported 8.4 percent; the CEC 8.9
PURSUIT: Physicians reported 8 percent; the CEC 14.2
As an example of how disagreements could occur, Mahaffey
cited the case of a patient who is rushed to an emergency room
with a heart attack and is enrolled in a trial. A few hours
later, the patient experiences some chest pain with some EKG
and blood enzyme changes.
"Did that patient have another heart attack? Not all
cardiologists would agree," Mahaffey said. "So we must come up
with ways we formulate our definitions to ensure that everyone
is following the same guidelines. We're now looking at changes
that can be applied worldwide in the trials we're now
While the outcomes of these trials are important to the
companies whose drugs are being tested, they have no influence
once the trial begins.
"The companies, of course, have input in the overall design
of the actual trial, but once it begins, CECs are completely
independent. As a matter of policy, CEC members are blinded to
the treatment assigned to each patient."