Proposal: A Better Strategy for Developing AIDS Drugs

It is widely agreed that what is most needed for better AIDS
treatment is better drugs -- and that the drug-development
strategies of the past have failed to provide them. If the
Concorde report at the Berlin conference does lead to a
movement among activists and others against AZT and the other
third-rate AIDS drugs now available, then the success and
value of this movement will centrally depend on the quality
of the alternative drug-development strategies which it can
advocate. It is not enough to only say what has been wrong so
far; we also need to suggest ways to fix it.

To contribute toward this thinking, we have outlined what we
believe would be the most important advance today in
conducting clinical trials and moving the results into
clinical practice. Details remain to be filled in, and
changes to the overall picture will be required. But this
effort needs the work of many people, so we decided to
present what we can suggest now, instead of waiting to
develop it as a more finished proposal.

Antivirals -- and Viral Activity Tests

Two major directions now for drug development are (1)
improved antivirals, especially practical treatments which
suppress HIV in chronically as well as acutely infected
cells, and (2) immune-system treatments, both to control the
virus indirectly, and to repair damage already done. Our
choice is to start with antivirals, because today they are
easier to test. It is much easier to tell if an antiviral is
doing what you want it to do, than to tell if an immune
therapy is doing something that will help in the long run.

Screening drugs to test for antiviral activity in humans has
long been a minor part of conventional drug development --
even several years ago, when only the crude p24 antigen test
was available for this purpose. AZT, ddI, etc. do pass the
antiviral activity test. What has been done is to select a
few volunteers who have a p24 antigen value high enough to
read on the tests then available, have them start using the
proposed antiviral, and watch their p24 antigen levels for a
few weeks to see if they are reduced.

The strategy we propose would:

(1) Make this kind of test for antiviral activity in humans
central to HIV drug development, not peripheral to it.

(2) Take advantage of the advances in viral testing which are
about to become available, such as sensitive tests for HIV
RNA, which are not commercially available but can be used by
researchers through special arrangements. Meanwhile, testing
could also be done with the newer p24 antigen test (called
"acid dissociated" or "ICD" p24) which is already in common
use in medical practice, at least in most states (it has been
difficult to obtain in New York, due to the way medical
laboratories are regulated there).

(3) Use these small, fast antiviral activity trials to test
many kinds of antivirals, not only nucleoside analogs or
other reverse transcriptase inhibitors. The same kind of
trial will work for drugs which act at different parts of the
viral life cycle -- and even for certain immune-based
therapies or nutritional interventions, if they might reduce
viral activity indirectly. It is especially important to test
drugs which have shown activity against HIV in laboratory
tests with chronically infected cells, where reverse-
transcriptase inhibitors are not effective.

(4) When a drug is known to be safe, and known to show
antiviral activity in humans, then it should be made
available as an option to physicians and patients, without
waiting for a large phase III trial to get definitive proof
of clinical benefit. In testing treatments for an infectious
disease, control of the causative organism is not a
"surrogate marker" which must itself be validated by years'-
long clinical trials before it can be used. It is, instead,
the central goal of therapy.

(5) Rapidly develop the improved viral-activity tests which
already exist in research laboratories, to make them
available to practicing physicians, who can then use them to
individualize their patients' therapies. Physicians would
continuously monitor the test results, and whenever they
found viral activity, try the various drugs developed as
above (drugs known to suppress viral activity in humans),
until they find a drug, dose, or combination which suppresses
it. Then they would keep monitoring, and change the treatment
regimen again if viral activity ever came back.

(6) With sensitive enough viral activity tests, this strategy
could be applied to early HIV treatment also, for example
when T-helper counts are over 500. It might or might not be
necessary to treat everybody. But it will be necessary to
monitor everybody, to start treatment as soon as each patient
need it.

Advantages

The conventional approach to testing AIDS drugs is
illustrated by the Concorde trial. A drug or combination is
selected for testing, and then it is applied in a standard
way, with as little individual variation as possible, to
randomly selected members of a group of patients. Others are
randomly selected for one or more comparison treatments. The
message from Concorde is that these trials must run for
several years, and must look for differences in survival or
major disease progression, as short-term results or T-helper
count changes are not reliable. After an immense effort
requiring hundreds of volunteers and lasting for years, a
single drug or one particular combination is either accepted
or rejected.

The alternative strategy proposed above tests new drugs and
combinations in weeks, not in years. Then, once safety can
also be assured, the new treatment could be made widely
available as an option to physicians. Once a drug is known to
be safe and known to have good antiviral activity in humans,
physicians should be allowed to weigh its advantages and
disadvantages, in comparison with those of other available
treatments, and use their medical judgment.

Another advantage of the proposed strategy is that it
produces individualized treatments, giving a range of options
which can be tested and refined for each patient.
Conventional HIV drug development, based on statistical
averages, tests a single treatment on everyone.

It is now widely accepted that much of the progression of HIV
disease is caused by worse viruses (drug-resistant viruses,
syncytia-inducing viruses, etc.), which arise from mutations
as HIV reproduces. Drugs like AZT block only the reverse
transcriptase step in the HIV life cycle; in chronically
infected cells, where this step has already taken place, the
virus remains active. Partially blocking the virus, as the
current drugs do, creates ideal conditions for drug-resistant
and other mutants to evolve.

What needs to be done is to block viral activity as much as
possible, so that there will be very little reproduction and
little chance for harmful mutants to be produced. To do this,
we need different kinds of drugs, better viral tests to make
sure the drugs are working in a particular patient, and a
greater number and diversity of treatment options, so that
when one treatment fails it can be replaced.

What can be done to put this kind of strategy into practice?
The first step is to talk about it, and refine the idea. As
wider consensus develops, trials will be designed and
conducted, one by one. This doesn't need to wait for any
central decision.