Statement of Nelson L. Michael, M.D., Ph.D

On 2 September 92, LTC Redfield delivered a major address on
the three year follow-up of the Phase I RV21 patients. At the end
of the address he emphasized that there was no statistically
significant trend, either up or down in the cohort of 26 evaluable
patients after an average of two years follow-up. He pointed out
repeatedly, both by slide and orally, that the Phase I data should
not be interpreted in relationship to efficacy because it was
obtained in a non-placebo controlled trial. Again, with caveats,
he stated that the lack of change in viral burden could be
interpreted as evidence of safety for the vaccine.
I was told by LTC Redfield on 14 September 92, that he had
been advised by COL Burke that a RV21 "safety" meeting was to be
held the following day at the HMJF headquarters building in
Rockville; LTC Redfield asked that LTC Birx and I attend the
meeting even though we had not been asked by COL Burke. Attendees
included USN personnel from the NNMC and San Diego, USAF personnel
from Wilford Hall, COL Burke, Mr. John Lowe (Executive Director of
the HMJF), COL Oster, MAJ Steven Johnson, MAJ Jeffrey Lennox, LTC
Brundage, Dr. McCarthy, Dr. Vahey, LTC Birx and me.
The meeting was opened by COL Burke who dwelled on the general
issue of safety; LTC Redfield then described the clinical safety
data. Many of the same points were covered at this meeting as were
covered at the investigators meeting in San Antonio in April 92.
Next LTC Birx discussed the immunologic data set that supported the
ongoing safety of RV21. COL Burke then raised the viral burden
issue. LTC Redfield stated that, working closely with Drs. Vahey
and McCarthy, he had completed the full analysis of the 26 RV21
Phase I patients prior to the Chantilly meeting and that he agreed
that the data showed no overall trend in viral burden for the
cohort over two years. He stated that his analysis presented at
Amsterdam had come to a different conclusion due to it's being
based of viral burden data from the 15 patients for whom he had
such data available at that time compared to a natural history
control group with CD4 Cell count corrected data and an arbitrary
one-half log increase threshold. He then stated that he had
learned from Dr. McCarthy that his use of a control group with the
Amsterdam data was not statistically ideal and he stated that in
the future he would not use statistical treatments for the Phase I
viral burden data. He also stated that he would submit Phase II
data to multiple statisticians prior to public presentations
We then moved into a discussion of the issue raised repeatedly
at the Chantilly meeting about whether gpl60 vaccination was
associated with transient bursts of viral replication. Dr.
McCarthy described several statistical nalyses he had performed
that showed that this was not the case.