Proposal: Small Trials for Screening Antiviral Combinations



[Note: In our last issue, AIDS TREATMENT NEWS published a
proposed monitoring program for getting reliable data from
the very first use of a new treatment. The proposal below is
an alternative approach to getting data early. It was
developed by Bill Bahlman, an activist with the Treatment and
Data Committee of ACT UP/New York, who is working with
members of the Inter-Company Collaboration for AIDS Drug
Development in the hope of getting the ICC to sponsor such
trials.]

Rationale: With the hundreds and soon to be thousands of
possible antiretroviral drug combinations, it is imperative
that we develop a process of rapid screening in vivo [in
people]. Past experience in drug studies, emerging
technologies, and new understanding of HIV pathogenesis make
it possible. We cannot afford to wait the many years it would
take to study only a small portion of the potentially life-
extending and life-saving combinations. We also need to
design studies to mirror what risks people with AIDS are
taking on their own. We need to provide some safety and
monitoring information for these people as well as to provide
a controlled setting for those who would rather do this sort
of experimenting in a clinical setting. These are individual
studies testing only one open label combination at a time.
New open label combination screening studies should be
started at a minimum of one per month.

Study size: 30 patients.

Inclusion/exclusion criteria: CD4 count range, to be
determined. There may be two parallel studies, one for
volunteers who are antiretroviral naive and have a CD4 count
over 300, the other for those who are antiretroviral
experienced and have a count under 300.

Also, a minimal viral burden copy number -- possibly over
15,000 copies of HIV RNA per milliliter of plasma -- will be
required.

Study drugs: Each study will test three, four, or more drugs
in combination. Any number of approved and/or unapproved
drugs can be used in combination together.

Length of study: As short as 6 weeks, no longer than 10 weeks
(to be determined).

Endpoints: Viral burden, CD4 and CD8 numbers, CD4 and CD8
percentages. Drug safety and toxicity.

Later research: Drug combinations providing a great viral
suppression and a great improvement in lymphocyte count and
percentage (to be determined) would rapidly be put into a
controlled phase III study with other such combinations.
These improvements in markers would also have to be seen in a
high percentage (75 to 90 percent, to be determined) of
patients. Volunteers should be given continued access to the
study regimen, open label, regardless of the study results.
During this continued access, they should be allowed to add
any other antiretroviral drug of their choosing, with
appropriate monitoring provided. The volunteers should not be
excluded from other drug combination screening studies.

Comment

The most controversial aspect of the above design is that it
proposes an uncontrolled trial, since there is no randomly
assigned control group to compare the test regimen against.

The ICC first planned to do its studies as uncontrolled
trials, but under pressure from some advocacy groups and from
the FDA, the ICC changed its design to a series of randomized
controlled studies. The change resulted in great delay, and
the resulting trials are beginning to recruit only now (see
announcement in AIDS TREATMENT NEWS #219, March 24, 1995).

The importance of having a randomized assignment to a
comparison group will not be known until the results are in;
only then will it be possible to see how much this comparison
adds to what we learn from the study. Meanwhile, we strongly
believe that non-randomized trials, such as the design
proposed above, do have a role, for several reasons.
Randomization adds greatly to the delay and difficulty in
setting up a trial and in recruiting volunteers; it is much
easier to get people to volunteer when they know exactly what
treatment they will get, than when they are told that a
computer will assign them to one of several different
regimens. Often there is no obvious treatment to use as a
control, and so one is chosen arbitrarily; in the ongoing ICC
study, for example, the control groups are those which
consist only of FDA-approved drugs, although it is unclear
why this comparison has any particular importance. Also, the
trials being discussed are screening studies, not definitive,
pivotal trials; combinations which look good in these early
studies will be entered into later trials which of course
will be randomized and controlled. Unless very minimal
improvements are being sought -- which is not the case here
-- randomization will not be required to show which
combinations are good enough to be looked at again in a more
formal study.

One addition we would make to the above proposal is to
specify followup visits going out for a least a year after
start of treatment, and perhaps indefinitely, with the
volunteers allowed to take or not take the study drugs as
they wish. The proposed six to ten week treatment period is
more than enough to show antiviral activity of a treatment
regimen. But we also need to know how long that activity
lasts, and what the long-term side effects may be.

Persons with comments on this exploratory trial design can contact
Bill Bahlman at 332 Bleecker St., Suite G-6, New York, NY 10014.