Expanded Access to Experimental Drugs: Interview with David Feigal, M. D., of the FDA

Editor's Note

David Feigal, M. D., is Director, Division of Antiviral Drug
Products of the U. S. Food and Drug Administration, in
charge of approving all new antiviral drugs. AIDS TREATMENT
NEWS asked for this interview for several reasons:

(1) There is much discussion among patients and physicians
about "when is a drug ready for parallel track." The FDA
must approve any parallel track or other such access program
before it can start. And (at least equally important in
practice), the FDA's views will be influential in
negotiations with pharmaceutical companies, which must also
agree to (and usually pay for) such access to their new
drugs. Understanding the FDA's approach will enable
physicians, patients, and activists to direct their energies
most effectively toward obtaining wider access to promising
treatments.

(2) The FDA wants help from the AIDS community in
identifying which experimental drugs are truly promising and
should be made more widely available. Physicians and
patients can help in this process.

(3) The FDA is prohibited from releasing proprietary
information about specific drugs. This puts AIDS activists
at a disadvantage in negotiating for access to treatments,
because companies can blame the FDA when the real obstacle
is their own reluctance; the FDA cannot defend itself, and
the public finds it very difficult discover what the truth
is. This problem can be minimized by documenting the fact
that the FDA is more supportive of pre-approval access to
experimental drugs than many people believe.

Glossary

A few of the terms used below will be unfamiliar to some
readers. These brief explanations may help.

* "The Agency." The U. S. Food and Drug Administration is
informally called "the Agency" in drug-development
discussions. There is no "Food and Drug Agency" -- although
this phrase often appears in the press, because of
reporters' confusion.

* Sponsor. An organization which is developing a drug (and
usually paying the large bills to do so, averaging well over
$100 million in development costs for each new drug
approved). Almost always the sponsor is a pharmaceutical
company. But occasionally a government agency or private
nonprofit organization can be a sponsor.

* IND (Investigational New Drug approval). Usually an IND
is permission granted by the FDA to test an experimental
drug in humans. In some cases an individual physician can
apply for and hold an IND. Certain special kinds of IND,
such as the "treatment IND," are explained in the interview
below.

* Phase I, phase II, phase III. These refer to traditional
phases of human trials of an experimental drug. Phase I
looks for toxicity and for the maximum tolerated dose.
Phase II seeks to show some indication of efficacy. Phase
III compares the new drug in larger trials against existing
treatments to determine its real-world usefulness.

These phases are not defined or even mentioned in any law or
regulation, but represent customary practices which
developed over the years. Therefore the FDA has flexibility
to allow new approaches. For example, in testing drugs for
AIDS and other life-threatening conditions, phases I and II,
or phases II and III, are sometimes combined to save time.

* NDA (New Drug Application). The NDA is a voluminous
document (typically dozens of boxes of paper) submitted by a
sponsor to the FDA to apply for approval to market a new
drug.

* Accelerated Approval. This is a new system by which the
FDA can grant full approval to an urgently needed new drug
based largely on "surrogate marker" endpoints (such as blood
tests), instead of waiting for proof of long-term clinical
benefit -- provided that the company agrees to do the
additional research (to prove clinical benefit) after
approval. The only AIDS example of accelerated approval so
far is the drug ddC; ddI was approved in a similar fashion,
only before the accelerated-approval regulations were
written.

* Expanded access. "Expanded access" is a generic term
which applies to a number of different programs to allow
patients to use an experimental drug before it receives
final marketing approval.

* Parallel track. This is the most recent expanded-access
program. The only official parallel-track program so far is
for the drug d4T. The major pre-approval access program for
ddI was in fact parallel track (but not officially so, since
the parallel-track regulations had not been written at that
time).

Interview with David Feigal, M. D.,

May 5, 1993

JJ: During an earlier conference call [on March 15], FDA
Commissioner David Kessler strongly supported parallel track as a
way to make important treatments available, so that patients do
not need to resort to underground treatments and use drugs
without quality control.

DF: Let me put parallel track in the context of the
different procedures which are available and have been used. It
is part of an overall commitment by the Agency to make drugs
available early to patients with life-threatening diseases.
Parallel track is based on the premise that if the drugs are far
enough along to test them in large-scale trials, then it is also
reasonable to make the drugs available to the patients who cannot
get into those trials -- particularly if they have no other
alternative.

If you look at the different regulatory mechanisms for this
purpose, one of the oldest is the "emergency IND," sometimes
called "compassionate use." We avoid the word "compassionate"
because during testing you don't know if the drug works or not,
so it is hard to tell if you are being compassionate or not,
depending on whether the drug is any good. Emergency INDs are
essentially single-patient protocols, where the sponsor has to be
willing to provide the drug, and the physician has to be willing
to take responsibility for describing the patient and how the
drug was used, and giving us some follow-up. These programs have
had as many as 300 or 400 patients per year enrolled in them;
some of the drugs have been used for HIV, usually for
opportunistic infections. These are particularly useful when a
sponsor is not pursuing or is barely pursuing licensure for a
drug, perhaps because the market in this country is too small;
yet the drug is available because it is approved somewhere else
in the world, and the company is willing to supply it. Sometimes
these programs have been coordinated by the Centers for Disease
Control, and they take on some of the sponsor's responsibility.
For example, they are now providing streptomycin and will soon be
providing PAS [both old drugs for tuberculosis, needed now
because of resistance to the standard drugs]. Their largest
program in AIDS was when they were the sole provider in the
country for pentamidine, for injection use for treating
pneumocystis.

The next kind of program under a traditional IND that
companies can sponsor is what we call "large safety trials."
These are not controlled trials; everybody gets drug. They only
differ from emergency INDs in that the sponsor takes
responsibility for the reporting and tracking of patients, rather
than individual physicians one-by-one as with an emergency IND.
Some of the safety data used for AIDS drug approvals have been
done under this mechanism of the large open safety trial. Some
of the early use of ddI, for example, used this method.

Historically, the next kind of availability that came along
was, for cancer, the "class C" drugs, and for non-cancer, the
"treatment IND." The treatment IND usually is used when a drug is
substantially complete in getting its phase III trials done;
typically it is filed about when an NDA is filed, or while an NDA
is being prepared, so that in the six months it may take the
company to prepare an NDA, and the six months it takes us to
review it, the drug can be available to patients.

Also, the treatment IND can be used earlier. For example,
the ddI expanded access program technically was a treatment IND.
But there are certain things in the treatment IND language [in
Federal regulations] that limit its use. For example, there has
to be no satisfactory alternative; and the manufacturer has to be
pursuing marketing approval with due diligence. "No satisfactory
alternative" creates a difficulty in certain conditions, for
example with early treatment, or combination use. If there is no
consensus that those are proved indications, the lack of
satisfactory alternatives doesn't ring true.

So for earliest use, when you don't know where the drug fits
in, parallel track can be helpful. Parallel track has some
features that are unique. One is that it specifies in the
regulations that the Parallel Track only applies to AIDS, whereas
all the other mechanisms are available for all life- threatening
diseases. Parallel track was based on the idea of an access
program that would run in parallel to the phase II or phase III
trials, more typically the phase III trials. It is for patients
who geographically could not get into those trials, or who did
not meet some of the entry criteria.

The only real parallel track so far has been d4T. But the
large programs done with ddI and ddC were in spirit parallel
track. Since the parallel track rules hadn't been written yet,
we borrowed other regulations.

JJ: Does "schedule C" [a program for cancer drugs] work
differently? There seem to be many more drugs to test with a few
patients, as part of serious trials.

DF: There are different mechanisms in cancer, partly
because of the way the programs have been set up. For example,
the National Cancer Institute (NCI) has a library of different
drugs that have been screened in tissue-culture systems. An
individual investigator who wants to get one of those drugs files
a letter of intent with the Investigational Drug Branch of the
NCI, and they can work out a protocol where they treat their own
patients, maybe one, maybe a small study of a group of patients.
What is different is that the NCI has a mass screening from big
chemical libraries.

They have candidate drugs that have gone through laboratory
screening and some animal toxicology. For drugs that belong to a
commercial sponsor, they have a standard way of setting up the
business arrangements between the government, the sponsor, and
the investigator. Historically, if you look back at what many
people envisioned for the ACTG [AIDS Clinical Trials Group of the
U. S. National Institute of Allergy and Infectious Diseases],
they thought that might be how the ACTG would be set up. But AZT
changed the environment to doing active-control, larger trials,
instead of having a network of physicians who would screen dozens
or hundreds of candidate drugs. It is a challenge to find the
best way to do the initial human studies; I think it's a mixture
of science, good intentions, and historical accident that we have
the system we do today.

JJ: How do we get around the existing obstacles in the AIDS
drug-development system? When there is a good rationale for a
treatment based on evidence, there may be no way physicians and
patients can choose the drug. Parallel track, etc., all require
that a sponsor be willing to participate. What incentive do
sponsors have to accept the cost, and possible liability
exposure, for no direct financial benefit?

DF: Let me comment from the FDA perspective first, about
what we consider a good candidate. What are the kinds of drugs
that prompt us to call the company and say you ought to be
thinking of some kind of expanded-access program?

One, there has to be some evidence of promise that the drug
will be effective. And almost more important at the early stage,
is that we need to know enough to be confident that it will be
safe. One of the assumptions we make with the expanded-access
mechanisms is that the drugs are going to be more loosely
supervised than they will be in clinical trials. Because of that
loose supervision we need to be sure that it's not toxic to the
liver, lymphocytes, etc., and would do more damage than good.

The evidence for effectiveness requires a case by case
decision. If the drug is in a class where we already have
approved drugs, there has to be more of a case for expanded
access than for a drug in an entirely new class [which might have
unique value]. Now we have a number of drugs in new classes,
such as tat inhibitors, protease inhibitors, and non-nucleoside
reverse transcriptase inhibitors, that are coming along in
testing. None of them have gone far enough to show the kind of
effectiveness that we would feel strongly for pushing for early
access. We're beginning to get data in. We hope that something
in these classes will be promising. But the data cannot just be
laboratory work with cell cultures, it cannot just be animal
effectiveness; there must be other evidence, usually from
clinical trial. For example, when the ddI access program started
there was data on about 100 patients.

It takes some evidence of effectiveness, and it has to meet
some need; there has to be some public-health need for expanded
access. For example, with pneumocystis treatment, trimethoprim-
sulfa is good enough that there's no reason to have an expanded
access for a first-line drug for pneumocystis. But, given the
allergic reactions with available drugs, and given the fact that
severe cases of pneumocystis are difficult to treat, it would be
fairly easy to make a public-health case for expanded access to a
new drug for treatment failures. It goes case by case, depending
on the disease, and depending on what other drugs are available
and what their strengths and weaknesses are.

You asked about commercial disincentives. With a couple
exceptions, most of the expanded access that has taken place has
been initiated by industry. There are companies that expressed
an interest very early. And note that the company that sponsored
the largest and probably most expensive expanded access, ddI,
came back for more when they developed d4T and wanted to do a
parallel-track program.

Why companies do expanded access is a question you could
best ask them. For the large companies, there are a number of
advantages to the program -- in learning about the safety of
their drug, and in getting visibility for a drug (especially
important in a crowded market). I think the main incentive for
them is that it cements in the public's eye that this is an up-
and-coming drug, one valuable enough that it's being made
available early. That kind of market influence is hard to come
by in other ways. The decision is probably a combination of
business interest, the chance to learn about the drug that way,
and in some cases a strong element of public service.

Companies can recover expenses of expanded access programs
by charging patients for drugs before they are approved, if they
have the FDA's permission. But this very seldom happens.
Companies fear that the documentation they would have to produce
to justify the price would reveal more about their business
practices than they wish to have known, since currently that
material is private and not reviewed by anybody. And many of
these drugs have a long development process; it's hard to tell
early how large a market they will have. So companies may be
concerned about setting a price for expanded access, as it may
not reflect the price they eventually want to charge when they
get to the marketplace.

Except for pricing of an experimental drug, the FDA
currently has no jurisdiction on pricing. Under the current law
it's none of our business. My guess is that the companies enjoy
that lack of supervision and don't want to get it started in any
area.

The economic disincentives of expanded access are definitely
a concern for the small companies. They have economic problems
everywhere they turn; drug development is an expensive business,
requiring animal testing, clinical trials, etc. This is one more
area where a small company will say they cannot afford a program.
The tougher position is sometimes for the bigger companies that
overall are large companies, but may look at some product areas
as though they were small companies. If they don't see them as
having large markets, they may not put much resources behind
them. Sometimes there are resource decisions not to pursue
expanded access.

I and others have speculated that one reason some companies
do early drug testing in Europe and not the United States is to
avoid the pressure to have expanded access after the first couple
of trials are approved. In Europe they can see if the drugs work
or not, without having to worry about expanded access. If these
trials are successful the company can begin their development in
the United States, and live with an expanded access as part of
the cost of that.

There are other reasons companies do trials in Europe
instead of the U. S. Many are European companies; they have
public-relations problems when they don't start their studies in
their own countries. Another reason is that it is easiest to
evaluate a new drug effect in patients who have never taken an
antiviral drug before. Because AZT is not used in early disease
in Europe, and ddI still is not approved in the United Kingdom
and many other European countries, it is easier to find drug-
naive patients in Europe.

JJ: When a small company has developed something, and for
whatever reason it has not been sold to a big company, are we out
of luck? The case of peptide T is on peoples' minds. What can
we do for the future? We don't see any solution in such a case.

DF: It is tough. Small companies have trouble even getting
past the 100-patient trials. As soon as they get into the
several-hundred-patient trials with longer-term follow-up, or
with complete viral or immunological monitoring, they may not
have enough money to proceed. They are stuck at that point with
having to find a business partner. They might be able to try it
with the ACTG; there are drugs from small sponsors that have had
some testing there. They need to find a source of backing, and
traditionally it is a partnership with a larger, old-style
company. That is a problem for them whether or not there is
parallel track.

One proposal has been to require more early animal testing
for reproductive toxicology, so we can make early access to drugs
a reality for pregnant women. If we took the cost of those
studies to the small companies before they had even screened the
drugs in phase I/II for efficacy, that would discourage the
development of some compounds as well. It's a rock and a hard
place, some of the issues you get caught between in drug
development.

What about interference with drug development by these
expanded access programs? The answer has been mixed. With AZT,
there was no effect, since the treatment IND was a bridging
maneuver during review and before full marketing. It did not
compete with any trial, and it provided drug to three or four
thousand patients half a year earlier than they otherwise would
have received it.

With ddI, the ACTG does not think that early access
interfered with recruiting for the trials. Concern about
pancreatitis had a chilling effect for several months. But
recruitment was nevertheless very close to what had been
targeted.

ddC did get clobbered, but not by its own expanded access,
but by ddI. ddI was the drug that everyone wanted to get on, and
it sort of shut everything else down. If there had not been ddI
expanded access, maybe the ddC clinical trials would have accrued
more rapidly, but only to the extent that patients on expanded
access would have gone to ACTG centers and signed up.

With drugs for opportunistic infections, most of the
expanded access has been after the trials were finished, with
the exception of atovaquone (Mepron). There was a fair amount of
atovaquone availability while trials were still going on. I
don't see any evidence that expanded access for atovaquone
interfered with recruitment for its clinical trials. There was a
lot of enthusiasm for that drug during its testing, and both the
clinical trials and the expanded access accrued fairly well.
Clarithromycin had an expanded-access program during its trials,
but the sponsor did not have trouble conducting the studies which
are now being presented (to the FDA's advisory committee).

The concern is a real one. It would be possible to set up
widespread use of drugs through expanded access or accelerated
approval that would have a discouraging effect on further
studies. Accelerated approval may be a much bigger threat than
parallel track for dampening enthusiasm for clinical trials. I'm
concerned at the difficulty there has been in launching
combination trials of ddC. Once a drug gets accelerated
approval, that makes its use so widespread that it is hard to
find participants who are interested, and investigators may want
to move on to new drugs. Unapproved drugs somehow seem more
attractive than approved drugs.

JJ: From here it looks like the problem is getting
companies motivated to do the trials when there has been
accelerated approval.

DF: We hope that the initiative in industry to be more
cooperative in combination trials will break down some of the
barriers. It often takes the companies a very long time to
decide whether to do studies with their competitors, even when
the trial provides no way for them to lose, since with
combinations, the patients will be taking both drugs.