Clinical Evaluation of Therapeutic Vaccines for HIV Seropositives: Approaches to the Design of Efficacy Trials

Consider the following scenario that could emerge from the traditional
approach:
Phase II study in CD4 < 200 shows no effect;
Phase II studies in CD4 > 200 show delay in CD4 decline.
According to the strategy described earlier, the implication of these results
would be the development of large studies in the CD4 > 200 categories to
assess definitively the clinical effect of the vaccine in these populations.
There is reason for concern, however, that an apparent positive effect on
markers may make it impossible to implement further placebo-controlled studies
that could establish clinical benefit. (The lack of effect in the CD4 < 200
group would likely not be overly discouraging since benefits in this group may
be hoped for but are not really expected.)
The most troublesome scenario that might emerge from the modified approach
would of course be that in which neither clinical benefit nor marker effects
were seen in any of the 3 patient categories. In this case, as many as 30,000
patients might have been treated with an ineffective (or even harmful) agent,
80% of whom would have avoided treatment had smaller marker-based studies been
performed in the CD4 > 200 categories prior to implementing large-scale
trials.
It is difficult to assess which approach minimizes the likelihood of a
potentially harmful vaccine product being administered to large numbers of
people. While the experience with therapeutic vaccines thus far suggests that
these products are quite safe, the theoretical possibility of harm cannot be
entirely discounted. A reasonably substantial effect that would be detected
by laboratory markers (e.g., an increase in viral load) might well be observed
in Phase II studies conducted according to the traditional strategy,
preventing the exposure of the thousands of patients who would have been
treated if the modified approach had been followed. On the other hand, an
adverse effect not well mediated through the markers might not be detectable
except in a large study of clinical effects. If the large, clinical endpoint
study were never performed (as suggested in the first scenario above), this
adverse effect might never be detected while the agent became standard therapy
for patients in the early stage of disease.
In summary, it seems clear that if vaccine-induced effects on CD4 counts
and/or other markers are reliable indicators of therapeutic effects, the
traditional approach is more efficient and more scientifically sensible. If
vaccine-induced effects on these markers do not reliably translate to clinical
effects, then the traditional approach may prove problematic.

12