AIDS TREATMENT NEWS #222, May 5, 1995
A major study of d4T (generic name stavudine, brand name
Zerit) has shown that patients who have been using AZT did
significantly better in disease progression if they switched
to d4T than if they stayed on AZT. This new information is
important for making treatment decisions; it confirms the
decision to grant accelerated approval to d4T last year.
Also, it shows that improvement in blood tests, including CD4
count and viral load in peripheral blood cells, can predict
later clinical benefit to patients.
The new information, from the '019' study sponsored by
Bristol-Myers Squibb, was presented at the Eighth
International Conference on Antiviral Research, April 23-28,
in Santa Fe, New Mexico. This study included 822 patients,
with CD4 (T-helper) counts between 50 and 500. All had taken
AZT for at least six months, but many had been taking it much
longer than that. Ten percent of the patients had a diagnosis
of AIDS at entry.
These 822 patients were randomly assigned to either continue
using AZT, or to switch to d4T; neither the patients nor
their physicians knew who was getting which drug. Patients
were on the study for at least two years. Three outcomes were
measured: (1) survival; (2) clinical progression (did the
person either die, or contract any new or recurrent AIDS-
defining opportunistic infection); and (3) treatment failure
(death, opportunistic infection, or a 50% or greater drop in
90% of those assigned to the d4T group survived for the two
years of the study, vs. 86% of those who stayed with AZT;
this is a clear trend in favor of d4T, but did not reach
statistical significance due to the small number of deaths.
The other two outcomes -- disease progression and treatment
failure -- were statistically significant in favor of d4T,
however. Switching to d4T resulted in an average three months
delay in clinical progression, compared to staying on AZT.
(While three months may not seem like much, in actual
practice the results are likely to be better than this group
average, because those who do best on d4T can stay on it,
while others can move on to other drugs or combination
treatments. This study was analyzed on an "intent to treat"
basis, meaning that those patients randomly assigned to d4T
were counted with the d4T averages, even if they could not
tolerate the drug -- therefore underestimating the treatment
effect.) Throughout the study, CD4 counts were higher by an
average of about 40 in those who switched to d4t.
The treatment groups were also analyzed separately by CD4
subset -- 50-100, 100-300, and 300-500. All three groups
showed a large "relative risk" in favor of d4T; these were
statistically significant in the 50-100 and 300-500 groups.
Viral load (HIV in peripheral blood cells) will be analyzed
in this study; blood was collected and plasma HIV RNA will be
measured. However, the tests have not been run yet; the only
virological data now available is the measurement of virus in
blood cells, which was presented to the FDA a year ago.
There were several differences in side effects between the
treatment groups, with d4T more likely to cause some side
effects, AZT more likely to cause others. Side effects are a
concern with d4T; we are hearing that a number of people are
unable to tolerate the drug.
This study does not answer a number of other questions about
d4T, including how well it may work as initial therapy.
D4T was approved by the FDA on June 27, 1994. This followed a
cautious recommendation for approval by the Antiviral Drugs
Advisory Committee, which met on May 20, 1994.
At the time of the May 20 meeting, data from 359 of the 822
patients was available; the other patients were still in the
blinded study, so their data could not be used. At that time,
those who were assigned to switch to d4T had an average CD4
count about 50 higher than those who continued AZT. The
amount of HIV in peripheral blood cells had dropped 53
percent for those on d4T, compared to an 11 percent increase
for those on AZT. (This measure of viral load is not the same
as the plasma HIV RNA which is more commonly used today,
although it is related.)
At the May 20 meeting, there were certain technical
controversies concerning whether d4T should be approved.
These controversies are outlined in "d4T (Stavudine):
Approval Recommended," AIDS TREATMENT NEWS #200, June 3,
1994. The main concern is that while blood work showed clear
improvement, there was not enough clinical data from the
trial yet to prove that those who switched to d4T did better
The FDA decided to approve d4T under its "Accelerated
Approval" regulations, meaning that it granted full approval,
but with the requirement that the '019' study be completed,
in order to confirm that the benefit seen in blood work would
be followed by actual benefit to patients. That confirmation
has now been presented.
It is likely that Bristol-Myers Squibb will now ask the FDA
to broaden the "labeling" of d4T, so that it will be
officially indicated for patients who have taken AZT for six
months or longer.
Besides its immediate relevance to drug therapy for HIV, this
trial strengthens the evidence that viral load reduction
caused by a drug is predictive of clinical benefit. Since
patients were assigned at random to one of two courses of
therapy (either switch to d4T, or remain on AZT), the
difference in viral load (virus in blood cells, in this case
-- not virus in blood plasma, which is easier to test and
therefore more commonly measured) was clearly caused by the
difference in treatment. The big question has been whether a
change in viral load caused by a drug will predict a change
in clinical outcome for the patient. In this case it did;
those in the treatment group which had the lower viral load
were also found, a year later, to have done better
More information will be available when the viral tests are
It also appears that a LARGE increase in CD4 (T-helper) count
as a result of drug treatment may be a much better predictor
of future clinical improvement than a small CD4 increase.
AIDS TREATMENT NEWS did not attend the conference in Santa
Fe, where these d4T results were presented. We obtained the
information from Ben Cheng of Project Inform, who is
preparing a more detailed article which will appear in the
next issue of PI PERSPECTIVES. You can receive that issue
when it comes out by joining the Project Inform mailing list;
to do so, call the Project Inform hotline at 800/822-7422
(415/558-9051 in San Francisco), Monday through Saturday10
a.m. to 4 p.m. Pacific time.
Project Inform has set a rare example of putting public
interest above organizational considerations in giving us
this information and allowing us to publish it before their
article comes out. What a contrast to the professional
journals which routinely delay important information for
months so that they can time its release for maximum
advantage to themselves.