Garnet Anderson, Ph.D.
Investigator and Biostatistician
WHI Clinical Coordinating
Member, Public Health Sciences Division
Hutchinson Cancer Research Center
July 9, 2002
Press Conference Remarks
Release of the Results of the Estrogen Plus Progestin
Trial of the Women's Health Initiative:
Data and Safety Monitoring
Thank you, Marcia.
At the Women's Health Initiative Clinical Coordinating
Center, we are responsible for managing, analyzing and reporting the data from
the more than 40 clinical research sites of the WHI. I am here to explain how
the data were monitored, leading us to the conclusions we are discussing
Throughout the estrogen plus progestin trial, we placed a
high priority on monitoring and maintaining participant safety, as well as on
ensuring that all WHI participants were well informed about any potential risks
associated with the study. One of the primary mechanisms for ensuring the
safety of research subjects in any clinical trial is oversight by an
independent Data and Safety Monitoring Board, or DSMB.
In accordance with NIH policy, members of the Data and
Safety Monitoring Board for WHI were appointed by the National Heart, Lung, and
Blood Institute. The group was composed of 12 individuals with expertise in
clinical trials, statistics, ethics and all of the relevant areas of medicine.
The DSMB met every six months to review the emerging
results. A plan established in advance was used to guide the board in
interpreting the data. The plan was set up to ensure that clear answers to the
primary questions about heart disease and breast cancer would be obtained,
while keeping open the possibility that important effects of estrogen plus
progestin on other diseases could emerge. Thus, hip fractures, pulmonary
embolism, stroke, endometrial cancer and colorectal cancer were all designated
as outcomes requiring monitoring, based on findings from previous studies of
hormones. To assure that serious effects not previously identified would not be
overlooked, deaths from all other causes were also monitored.
To weigh the observed risks and benefits to the
participants' health, the DSMB used a "global index," which summed all of these
designated health outcomes.
The monitoring plan allowed for a variety of possible
scenarios and indicated how the DSMB would evaluate the data in each of these
circumstances. For example, if the data showed that estrogen plus progestin
helped prevent heart disease, and the global index indicated overall benefit,
then the DSMB would consider recommending that the study be stopped because the
benefits of hormones would have been adequately determined. Conversely, if
breast cancer or other risks became clear and the global index pointed to
overall harm, the board would consider recommending that the study be stopped
for adverse effects. For all other scenarios, the study would continue as
planned because the balance of risks and benefits would not have been
So, here is how the estrogen plus progestin study unfolded.
First, in late 1999, with an average of two and a half years' worth of data on
the trial participants, the DSMB noted that the rates of coronary heart
disease, strokes and pulmonary embolism, or blood clots, were higher among the
women taking the combined hormones than among those taking placebo. The DSMB
recommended that participants be told these interim findings. The WHI
investigators did so in the early spring of 2000 and again in spring 2001when
it became clear that the risk persisted beyond the first few years.
These results did not meet the established criteria for
stopping the trial, however, and they were not strong enough to rule out the
possibility of longer-term benefits for either these cardiovascular conditions
or for other health concerns. In addition, no other significant risks had been
identified at that time. Therefore, the DSMB recommended that the trial
continue with no changes to the protocol.
More recently, during its May 31st meeting, the DSMB
reviewed WHI data through February 2002. Then, with an average of 5 years of
participant follow-up, the increased risk for heart disease, strokes, and blood
clots was still apparent. The DSMB also discovered for the first time, that
breast cancer rates were markedly increased among women assigned to the
estrogen plus progestin group. Because this effect emerged suddenly, we wanted
to be sure that it was not a random or temporary fluke in the data. We looked
at additional data accumulated through April 30th and found that the picture
had not changed. Compared to women on placebo, more women on estrogen plus
progestin had been diagnosed with coronary heart disease, stroke, blood clots
AND breast cancer.
There were some benefits to the use of estrogen plus
progestin--women on these hormones had fewer hip fractures and colon
cancers--but the magnitude of these benefits was not sufficient to overcome the
accumulated risks observed for other diseases.
The increased risk of invasive breast cancer among women
taking estrogen plus progestin surpassed the disease-specific criterion defined
in the monitoring plan for stopping the study because of harm. The global index
also supported a finding of overall harm, satisfying the second criterion of
the monitoring plan. Based on these results, the DSMB concluded that estrogen
plus progestin produced harmful effects on women's health that exceeded any
benefits generated by these hormones and unanimously recommended to the NHLBI
that the estrogen plus progestin study be stopped.
NHLBI followed the board's recommendation and began
informing the relevant parties. The Institute has briefed the WHI
Investigators, and they are working collaboratively to explain the findings to
study participants. All WHI participants, even those who are not in the
estrogen plus progestin study, are being sent letters about the results. In
addition, women in the estrogen plus progestin study are being instructed to
stop taking their hormone pills immediately, and staff at each WHI clinical
center are scheduling counseling sessions with each participant. These
participants are being asked to continue their clinic visits so that their
safety and health outcomes can be monitored.