ddI and ddC Approval Effort -- Interview with Martin Delaney

Martin Delaney, co-founder of Project Inform, has been
actively involved since August in the effort to obtain rapid
evaluation and, hopefully, approval of ddI and ddC. We asked him
to outline what is happening in this effort.

Before working full time in AIDS, Mr. Delaney taught
negotiation and other business skills to corporate teams.
Recently he has communicated extensively with the researchers
studying ddI and ddC, and with other experts from the sponsoring
pharmaceutical companies, the FDA, the National Institutes of
Health, and activist organizations.

JJ: You are optimistic on early approval. What are the
reasons for optimism?

MD: I have had extensive meetings with the parties -- the
FDA, the researchers, the companies, and other activists. There
are some uncertainties about how the data will be evaluated, but
there is a clear commitment on the FDA's part in making this
process happen. The discussions are about how they can make it
happen, not whether. The question is, how can we evaluate the
data in this circumstance, and make a scientific case that what
we are seeing is predictive of the usefulness of the drug?

The issues are not unique to ddI or ddC. These are broader
issues for all AIDS drugs in the near future. We will not have
placebo studies, so we must find ways to make comparisons to
existing AZT data, and to a better picture of the recent natural
history of AIDS since the beginning of pneumocystis prophylaxis.
If these issues did not come up now for these drugs, they would
come up for whatever drugs are next. There are scientific
hurdles, but there is also a strong commitment to clear these
hurdles, to find a scientifically valid way to do it.

There are no simple answers here. The simple answers have
already been rejected by FDA. For example, they do not feel that
T-helper increases -- even for a longer period than produced by
AZT -- is, by itself, adequate proof of the drug, unless it is
corroborated by other data. But others hotly dispute this view,
arguing that T-helper increases alone should be enough.

The FDA does not want to be the obstacle; it recognizes the
critical urgency of making the two drugs available. So they are
saying that they want a joint decision, a collective decision-
making process to explore all these issues, with scientists in
other research and development agencies -- for example, the
National Cancer Institute, and the AIDS Clinical Trials Group of
the National Institute of Allergy and Infectious Diseases, as
well as activist experts -- on how the NDA (New Drug Application,
with the detailed data on the drug) should be submitted. The FDA
has agreed that it is safer politically to seek expert consensus
in this way; if it made the decision alone without consultation,
it would be highly vulnerable if something went wrong. In the
FDA's view, this decision is a watershed event. We do not want
them to make it alone.

JJ: The biggest questions seem to revolve around what
constitutes proof of efficacy.

MD: Absolutely. The debate will hinge partly around the
meaning of drug-induced T-helper increases. Statisticians are
finding that T-helper counts definitely associate with survival
outcome -- especially that the risk is greater when the count
falls below 50.

The tougher question is, if the T-helper counts are so
predictive, does increasing the counts by use of a drug predict
increased survival, as a higher count does in the natural history
(without treatment)? Most researchers instinctively believe it
does [that an antiviral which increases T-helper count probably
does improve survival]. But the FDA argues that no one has yet
proven this point.

Recent statistical analysis suggested that the benefits of
AZT are greater than what would be predicted simply by the T-
helper increase. Why that is so is unclear. The T-helper count
may not tell us all we need to know about the immune system; the
benefit of AZT may be greater than this count alone would
suggest. None the less, others argue that increased T-helper
counts are one useful predictor of a drug's efficacy.

The National Cancer Institute, for example, says that in all
its studies, any antiviral drug which has kept the T-helper count
above 50 seems to keep people alive. In the NCI studies, only
one patient who remained above 50 has died in the last four
years. Samuel Broder, M. D., director of the NCI, strongly
believes that this data alone validates the usefulness of drug-
induced T- helper increases [as an indicator of clinical or
survival benefit from the drug].

The long-term clinical picture, as measured by the number of
opportunistic infections or by survival, is harder to evaluate
than people would like, because the easiest point of comparison
is the old AZT studies. And most of those were done without
pneumocystis prophylaxis. So the FDA is reluctant to use them
for comparison, arguing that more recent data includes other
variables besides the drug. So now Bristol-Myers is collecting
data from several major clinical centers in the country -- not
just ddI data, but also case-history data from patients not given
ddI, to get a better picture of the natural history of AIDS with
pneumocystis prophylaxis. Perhaps this data can be used as a
valid comparison; but of course much work is required.

Another outcome which FDA will consider will be overall
clinical improvement, such as weight gain, or Karnofsky scores (a
rating of overall health). If you put each of these pieces
together, and if they are coherent, I think there is no question
that FDA will grant the approval. But they do not want to
approve the drugs on T-helper counts alone, without this other
data.

There should also be some input from virology markers.
Plasma viremia data (measuring the amount of virus in the blood)
should be available by January.

If you keep your eye on all of these pieces of the picture,
you don't get nervous with every new rumor that sweeps part of
the country.

JJ: What else is at stake here, beyond the short-term
availability of ddI and ddC?

MD: A great deal is at stake. If the FDA cannot find a way
to accept this kind of data as proof of drug effectiveness, it
will have no choice but to demand the re-use of survival as the
sole endpoint in future studies. In other words, we test the
drug until some of the patients die. And this would require a
return to the routine use of placebo controls. We cannot allow
this to happen.

JJ: If a drug is known to be an antiviral, known to be
effective against HIV at concentrations reached in the body --
and if there is no reason to suspect that the drug raises T-
helper counts directly -- then if T-helper counts consistently go
up when the drug is given to persons with HIV, it seems hard to
explain that effect other than by an antiviral action of the
drug. Average T-helper cell counts do not rise spontaneously,
without treatment, in any known group of persons with AIDS or
HIV. It is hard for us to understand waiting months or years for
conclusive proof of benefit, after it is already clear that a
drug does show a substantial antiviral efficacy in patients.

Usually the FDA's Antiviral Advisory Committee meets for
only a day or two, with little or no staff support. How can it
possibly give the ddI and ddC decisions the attention they
deserve, in view of the extensive data about these drugs, and the
major issues regarding interpretation of the data?

MD: The committee probably could not do this by itself.
Therefore, it will be assisted by other leading researchers
brought in as consultants, who have expertise on these particular
drugs. A special meeting of the nation's top experts, probably
in early February, will seek a consensus on whether these drugs
should be approved. In practice, this process must go through
the existing structure of the FDA advisory committees.

JJ: There has been concern that neither company (Bristol-
Myers for ddI, or Hoffmann-La Roche for ddC) has yet submitted
its NDA (New Drug Approval application) to the FDA.

MD: The companies do seem to be on schedule for submission
of that data. It would be a mistake to submit it too early,
before they knew what the FDA wanted of them.

I have reassurances that Bristol-Myers will be submitting
its NDA shortly after the advisory committee meeting, probably in
February. Hoffmann-La Roche has not yet looked as closely at its
data, but they have said that they intend to submit their NDA
early in 1991.

JJ: What timetable do you see for a decision on approval of
these drugs?

MD: The advisory meeting date has not yet been set, but
should be in early February. I see at least one of the NDA
applications coming in about a week after that time. Then it
would take at least 30 days for the FDA to analyze the data; they
may have to ask more questions of the company. All things
considered, I think it's a do-able target to get both drugs out
by March of 1991. Anything much beyond that, and we should raise
the temperature politically.

The consensus building is most important, because it will
set a standard of how we look at these urgent approvals in the
future. We do not ask the FDA to decide all by itself in an
ivory tower. Instead, we want a collective decision with input
from all the people who are working in this field.

Note: On December 19, a multi-city press conference will
explain the movement for early ddI and ddC evaluation. Project
Inform and other organizations involved in this issue will take
part.

As of this week, 35 organizations have signed a consensus
statement circulated by Project Inform (with much help from ACT
UP/Golden Gate and Mobilization Against AIDS), calling for urgent
review of ddI and ddC. Also, the Community Consortium, a medical
group representing almost all of the physicians who have an HIV
practices in San Francisco, issued a separate statement urging
expedited review of the data on these drugs, with a decision on
licensing as soon as possible. In addition, at least 25 members
of Congress have signed a letter to the FDA by Congresswoman
Barbara Boxer (D-Greenbrae, CA) urging immediate review of ddI
and ddC.