June 2007 Pharmacy & Therapeutics Journal
When the Food and Drug Administration (FDA) approved
Merck’s quadrivalent human papillomavirus vaccine (Gardasil)
on June 8, 2006, the agency extracted some post-approval commitments
from the company, as it frequently does when it
“green-lights” an important new drug around which questions
still revolve. Those commitments involved additional testing of
Gardasil after it went on the market, particularly in
girls 11 to 12 years of age. Only a small number of
girls in this age group had participated in Merck’s
clinical trials, and these trials were submitted as
evidence of the drug’s efficacy and safety. That
patient population would loom large in the months
after the FDA’s approval as Merck mounted a political
campaign to convince states to require the vaccination
for sixth-grade girls.
However, the post-approval studies that Merck
agreed to perform would not start before June 2009,
if then. That short-term safety surveillance study is
set up to follow 44,000 vaccinated subjects for just 60 days to
monitor visits to hospital emergency departments and for six
months to monitor chronic problems such as autoimmune
disorders, rheumatologic conditions, and thyroiditis. Merck
must include a “sufficient” number of children (ages 11 to 12),
although no number is specified in the FDA’s charge to the
company.
Other companies that have committed to similar postapproval
studies have dragged their feet and, in some instances,
have simply ignored the FDA. The FDA has no authority to force
a company to complete a study. Even when a study is finished
and the FDA receives the data, there is no guarantee that the
results—if they are in any way disquieting—will prompt a suitable
reaction on the part of the agency, such as immediately
requiring new labeling or, in more serious instances, pulling the
drug from the market.
In the past few years, a number of unfortunate instances, particularly
the rofecoxib (Vioxx, Merck) fiasco, have shown that
it is easy for the FDA, because of shortcomings in the data, as
well as because of bureaucratic malaise and communication
gaps, to miss credible warning signs of drugs “gone wild,” producing
numerous adverse reactions.
During the past couple of years, these and other potential
problems have been highlighted, first by the U.S. Government
Accountability Office (GAO) and later by the Institute of Medicine
(IOM), whose report was requested by the FDA.1 Both the
IOM and the GAO cite the failure of drug companies to complete
post-marketing studies on time; they also cite the FDA’s lack of
authority to compel timely completion.
Serious limitations also afflict the FDA’s MedWatch system,
which catalogues reports of adverse drug events (ADEs). Moreover,
even when troublesome signs appear after a drug is on the
market, that evidence might not be available to the FDA’s Office
of Surveillance and Epidemiology (OSE), formerly
known as the Office of Drug Safety (ODS). The
OSE is set up to ensure that post-marketing problems
receive quick attention. It is subservient to
the FDA’s Office of New Drugs (OND), which has
considerably more stature within the Center for
Drug Evaluation and Research (CDER) and which
has sometimes been criticized for turning a blind
eye to emerging evidence of late-blooming safety
questions about new drugs.
Sheila Burke, MPA, RN, Chair of the IOM committee
whose members wrote its report, says:2
We found an imbalance in the regulatory attention and resources
available before and after approval. Staff and resources devoted to preapproval
functions are substantially greater. Regulatory authority
that is well defined and robust before approval diminishes after a drug
is introduced to the market. Few high-quality studies are conducted
after approval, and the data are generally quite limited.
These weaknesses explain why post-marketing surveillance
of drugs is the marquee issue as a major FDA reform bill sashays
(or, more appropriately, staggers) down the red carpet of the
Congress to the White House—because senators tussled over
multiple amendments when that bill, the Prescription Drug
User Fee Act (PDUFA) Reauthorization (S. 1082), came to the
Senate floor this past May. Controversial amendments involving
drug importation and the approval of follow-on copies of biotechnology
drugs were removed from the PDUFA bill. However,
although there were pitched battles over the provisions in
another amendment, the FDA Revitalization Act, sponsored by
Senator Edward Kennedy (D-Mass.), there was no question
that the post-marketing reforms in the Kennedy bill would be
included in the final PDUFA bill, and they were. However, those
post-marketing changes were not nearly as far-reaching as
Senator Charles Grassley (R-Iowa) had wanted; he had also
sponsored a competing FDA post-marketing reform bill.
The House of Representatives is likely to adopt the provisions
of the Senate’s bill on post-marketing surveillance, and the
bill will probably be sent on to President Bush to sign; he says
that his signature will be forthcoming. But before we look at
these changes in the Senate bill, it is necessary to outline the
problems they aim to resolve.
The FDA’s MedWatch system is the agency’s first line of
defense against a newly approved drug whose side effects,
whether or not they have been acknowledged in clinical trials,
cast a much darker shadow than previously thought. Physicians,
pharmacists, patients, and drug companies send in
reports of adverse reactions as the drug begins to be used and
potential problems become evident. But even the FDA has
admitted the system’s shortcomings. In a November 2006
speech, Scott Gottlieb, MD, Deputy Commissioner for Medical
and Scientific affairs at the FDA, acknowledged that MedWatch
is a valuable tool but “we know that it is not being used as effectively
as it ought to be.”3
The FDA and other agencies go by this general rule: for
every 10 adverse reactions reported, only one report is submitted
to the FDA. Of course, even that minimal number of reports
can be useful.
An example is Sanofi-Pasteur’s Menactra, a vaccine designed
to prevent bacterial meningitis. The FDA approved that vaccine
on the basis of clinical trials involving 7,000 children, none
of whom had contracted Guillain-Barré syndrome (GBS), a
neurological disorder. But lo and behold, when the drug reached
the market, reports of GBS developing in the children after
inoculation with Menactra began to filter in to the Vaccine
Adverse Events Reporting System (VAERS). VAERS is the
Center for Biologics Evaluation and Research’s version of
MedWatch.
Initially, there were only five reports. The FDA decided not
to change Menactra’s labeling, because those five reactions
were expected to be the norm with the introduction of a new vaccine.
At the time, however, the FDA noted that ADEs associated
with the vaccine were not always being reported and that there
might be additional cases that the agency might not be aware
of.
When Caroline Loew, PhD, Senior Vice President of Scientific
and Regulatory Affairs at the Pharmaceutical Research and
Manufacturers of America (PhRMA), appeared before a House
committee in May 2007, she gave a more expansive analysis of
MedWatch’s flaws:4
One of the shortcomings of this system is the variable nature of
reporting and the quality of reports. Ultimately, any database is only
as good as the underlying data, and one of the chief difficulties with
adverse event report databases is quality. Precious resources are
often expended in contacting health care professionals regarding
aspects of a report they have filed. In many instances, the reporter
is unable or unwilling to provide sufficient detail for analysis.
The FDA has been moving on its own to address this problem.
At the start of 2007, the agency announced steps it would
take with the extra $90 million it expects to receive in fiscal 2008
from the higher PDUFA fees that Congress is in the process of
authorizing.
One step would be to publish a request for proposals from
outside research organizations that would assist in determining
the best way to maximize the public health benefits associated
with collecting and reporting serious and non-serious ADEs
throughout a product’s life cycle.
Central to addressing this question are:
• determining the number and type of safety concerns
discovered by collecting the reports of ADEs.
• the age of products at the time safety concerns are detected
by the collection of ADEs.
• the types of actions that are taken next to protect patient
safety.
In early March, the agency took the next step by sponsoring
a public meeting to explore opportunities for linking postmarketing
monitoring systems in the private and public sectors
to create a virtual integrated, interoperable nationwide medical
product safety network. Such a “sentinel network” could integrate
existing and planned private and public sector databases
to enable the collection, analysis, and dissemination of safety
information about medical products to health care professionals
and patients at the point of care; that would be in the
clinic, where this information is needed to make informed
decisions about safe and effective treatments.
One of the Kennedy provisions in the FDA Revitalization Act
essentially seconds the FDA’s motion toward a sentinel network.
It authorizes a public and private partnership to establish
a routine active monitoring system that would create a “pool” of
relevant data assembled from federal and private electronic
databases of the health care population. This apparently would
be a complement to—or even a rival to—MedWatch. According
to the Kennedy bill, the data pool would have to have 25 million
patients swimming it by January 1, 2009, and 100 million patients
by January 1, 2012.
Although Senator Judd Gregg (R-N.H.) backed a competing
bill called the Safer Data bill, he agrees with Senator Kennedy’s
challenge to the FDA that it set up an active surveillance system.
However, he claims that the Kennedy timetable is too leisurely:5
I strongly support this in concept but feel the language needs to be
strengthened to ensure that the FDA has the direction it needs to
implement a robust system in an expedited timeframe. Information
collected must be standardized, and the overall system should be
validated.
Despite Senator Gregg’s criticism, it is difficult to see how the
FDA could assemble a pool of 25 million electronic health
records (EHRs) on Kennedy’s timetable, much less on a faster
one. There are 85 million Medicare and Medicaid recipients, but
neither of those two federal systems has mandated that physicians
produce EHRs. In the private sector, Kaiser Permanente
has been the leader in developing EHRs, and it now has them
for half of its 8.5 million patients, according to Paul Wallace, MD,
Medical Director of Health and Productivity Management Programs
at the organization.
The Kennedy proposal, which the House will probably accept,
does nothing to correct MedWatch’s debilities. It simply layers
a new early warning system based on EHRs on top of Med-
Watch. However, a second Kennedy proposal promises a better
benefit in the near term.
Another element in the Kennedy-sponsored FDA Revitalization
Act allows the FDA to order drug companies to prepare Risk
Evaluation and Mitigation Strategies (REMSs) for new drugs;
this might go a long way toward completing post-marketing
studies more quickly.
There is no question that drug companies have fiddled with
completing these studies, which explains why nearly everyone
in Congress is livid. The FDA itself has acknowledged that drug
companies decline to perform a significant percentage of the
post-marketing studies requested.
The agency’s own statistics for 2006 show that drug companies
failed to initiate more than 70% of post-marketing studies
that they committed to performing during that year; 899 of
the 1,259 post-marketing studies (71%) promised had not begun
as of September 30, 2006. This was an increase of 5% over fiscal
year 2005. When that report first came out, Representative Rosa
DeLauro (D-Conn.), Chairman of the House Appropriations
subcommittee, which approves the FDA’s annual budget, said:6
This report clearly demonstrates that drug companies do not intend
to keep these promises and that drug companies are taking advantage
of FDA’s lack of authority to require these studies.
However, PhRMA’s Dr. Loew has a different take on the
numbers:4
While it is true that 71% of open commitments are considered ‘pending,’
these ‘pending’ studies are in the preparatory phase of clinical
trial development during which the protocol is drafted and submitted
to FDA, IRB approval is obtained and the sponsor begins
recruiting clinical investigators. If sponsors simply failed to initiate
such studies, the studies would be coded as ‘delayed’ rather than
‘pending.’ However, only 3% of open studies are considered to be
‘delayed.’
No one disputes the value of post-marketing studies. Nothing
bears out their worth better than the pediatric studies performed
by drug companies in order to get an extra six months
of marketing exclusivity, which the Best Pharmaceuticals for
Children Act (BPCA) grants them.
For example, a 2007 GAO report about this act showed that
about 87% of the drugs that were granted pediatric exclusivity
under BPCA required labeling changes—often because the pediatric
drug studies found that children might have been exposed
to ineffective drugs, ineffective dosing, overdosing, or previously
unknown side effects.
And no one disputes the fact that the FDA does not have the
authority to compel drug companies to complete studies. The
Kennedy bill changes that by allowing the FDA to force companies
to accept REMSs when it approves a new drug. A REMS
could require that a company conduct post-approval studies, such
as a prospective or retrospective observational study, or even a
clinical trial, if a study was thought to be an insufficient means of
assessing a signal of a serious risk or of identifying an unexpected
serious risk. In addition, for the first time, the FDA would
be able to issue civil fines of up to $250,000 for not meeting a study
deadline, with the amount of the penalty doubling every 30 days,
up to a total of $2 million—figures that Senator Grassley insisted
be increased from the $10,000 and the $1 million levels in
the Kennedy bill. The Senate agreed by a strong majority.
However, by a narrow vote of 46–47, Senator Grassley failed
to establish a separate safety office within the CDER that would
have equal standing to the OND. Both the IOM and the GAO
highlighted the institutional weakness of the OSE, which at the
time of the GAO report was called the Office of Drug Safety
(ODS).
The GAO report from March 2006 described the OSE as subservient
to the OND. When the GAO finished its report, the
office had had eight different directors in the previous 10 years.
The GAO said that the ODS’s Division of Drug Risk Evaluation
was responsible for culling data on adverse drug reactions from
MedWatch and for making recommendations to the OND. However,
the “new drug folks” did not always listen closely to that
advice, and they even barred ODS staff members from appearing
before FDA advisory committees when those committees
were meeting to recommend the approval of new drugs and handle
other matters.
Equally disconcerting was another GAO finding: even when
the ODS made recommendations to the OND, no one at the
ODS stayed abreast of that particular problem to see how events
played out.
Steven Galson, MD, MPH, Director of the CDER, has
attempted to at least partially correct those bureaucratic disconnects
by elevating the OSE to report directly to him, instead
of through the OND. In addition, he established a new position
of Associate Center Director for Safety Policy and Communication
to focus on developing and implementing broad drug safety
and communication policies. Senator Grassley wanted to go
further by setting up a separate drug safety office on a par with
the OND, but the Senate rejected that approach. The Kennedy
bill does not address the OND/OSE imbalance at all.
However, a second Kennedy bill, called the Enhancing Drug
Safety and Innovation Act, which was also included in the
PDUFA reauthorization, attempts to broaden the flow of clinical
trial data into the FDA’s hands. It sets up a new national FDAadministered
clinical trials registry and requires that all clinical
trials supporting applications for drug approval, as well as all clinical
trials conducted with federal funding, be included in that registry.
When Senator Kennedy introduced this act on the Senate
floor, he explained the need for this new FDA registry by referring
to results from clinical trials on selective serotonin reuptake
inhibitor (SSRI) antidepressants, which have been linked to
suicide in teenagers.
He noted, “Tragically, such information was not adequately
available.”
However, all sorts of caveats are in the bill, allowing the Secretary
of Health and Human Services (DHHS) to waive reporting
requirements. (The FDA is part of the U.S. DHHS.)
Moreover, although it is difficult to argue with the [rationale]
for requiring quicker publication of clinical trial data, it is important
to keep in mind the limitations of that data. When she
appeared before the House Energy and Commerce Committee
on May 9, Marcia Crosse, Director of Health Care Issues for the
GAO, explained:7
Clinical trials typically have too few enrolled patients to detect serious
adverse events associated with a drug that occur relatively infrequently
in the population being studied. They are usually carried
out on homogen[e]ous populations of patients that will actually take
the drugs. For example, they do not often include those who have
other medical problems or take other medications. In addition, clinical
trials are often too short in duration to identify adverse events that
may occur only after long use of the drug. This is particularly impor-
tant for drugs used to treat chronic conditions where patients are
taking the medications for the long term.
Given the major problems that the FDA has had in recognizing
the initial warning signs of questionable safety and efficacy
associated with some new drugs, maybe it is too much to ask
that Congress fix those problems in one fell swoop. Certainly,
giving the FDA the authority to fine companies who miss deadlines
for post-marketing studies gives the agency important
new leverage. And the clinical trials registry can’t hurt. But to
the extent that the OSE remains a weak cousin within the agency
and to the extent that the MedWatch system remains as leaky
as a New Orleans levee in the wake of Hurricane Katrina, it is
likely that a new Vioxx-type situation will occur, perhaps sooner
rather than later.
