Highlighted Responses FOIA Request of October 4, 1999 concerning approval and
recommendation of rotavirus vaccine.
Indented items are direct quotations; a few spelling errors have been corrected
Lynn Armstrong, CDC/ATSDR FOIA officer
We are withholding predecisional internal communications. Release of this type of
material would interfere with the agency's deliberative process.
to MW from Mei-Shang 5/27/99
Intussusception was found in two cases of vaccinated children with a WC3 strain
trial in Shanghai. Result of the trial was reported in one of the Rotavirus Vaccine
meeting. Roger should have the information. (I believe it was the third meeting held in
Washington in the year 1990).
We learn about the intussusception when we were well into the trial. If one were to
calculate the rate it would have been very high because the trial we did was a relatively
small one.
from MW to Walt Orenstein, 5/27/99
According to Nelson's pediatrics book, the background rate for intussusception is 1
to 4 per 1,000 live births.
Given the occurrence of intussusception among recipients of rotavirus vaccine (1) in the
Shanghai trial; (2) in prelicensure trials of rotavirus vaccine in the U.S.; and (3) post-
licensure as reported to VAERS, we are quite concerned.
From MW to Trudy Murphy 5/28/99
I talked to Peggy Rennels. She wasn't too surprised, and said that she had TOLD
Peter Paradiso she thought there was something there pre-licensure. She also said that she
gave a talk at Emory (!) and that someone had come up to her and said that they had had a
case of intussusception post-rotavirus vaccine....
From MW to Bill Gallo 6/17/99
On June 17, 1999, CDC briefed the Advisory Committee on Immunization
Practices about cases of bowel obstruction following use of rotavirus vaccine in
infants....[12 cases reported to VAERS] No changes have been made in recommendations
for use of rotavirus vaccine.
MW to Larry Anderson 6/17/99
I could quibble with you about details (the VAERS data are alarming for a number
of reasons, and given the known problems with completeness of reporting to passive
surveillance systems, are consistent with an incidence markedly in excess to that expected
by chance alone) but I don't disagree with your bottom line concern. That said, it's
anybody's guess how this will turn out. We'll see what happens this afternoon. Even if
ACIP supports the current plan of action (continuing the vaccine while we collect more
data), I'm not sure what will happen if we get another 50 reports to VAERS next week,
after this hits the press. At the moment, we're taking this one day at a time.
From MW to Peter A. Patriarca (FDA) 6/21/99
Thanks for the comments. The intent was not to be dismissive (we are quite
concerned about this, and think that the reported cases are the "tip of the iceberg") but I
guess that got lost. Originally it was "substantially" higher but that got changed to
"somewhat" which I know doesn't say much.
MW to John Livengood 6/22/99
VAERS cases: I talked to Dr. Richard Ricketts, Chief of Pediatric Surgery at
Emory, this morning. He agreed to review the case summaries and after doing so called
me right back. What struck him about the cases was that they were "too young" (in his
experience, intussusception is very unusual among those <6 months of age) and the high
proportion that needed surgical correction (at Emory, only [about] 25% go to surgery)....
Threshold for public health action ("n") [or when do "anecdotes" become a series -- Ed.]:
In your abundant spare time there in Dallas, maybe you and Walt and Jose can decide
what "n" is for getting the ACIP on the phone for a conference call to reconsider the
recommendation for this vaccine. You don't have to tell me what it is, but I would
appreciate it if someone was talking about it.
MW to John Livengood, 6/23/99 re: expected rates of intussusception among RRV-TV
recipients, NCK
I was going over this with Mike Blum on the phone (safety guy with Wyeth
Lederle) and realized that person-time is wrong (infants entering the vaccinated cohort
REPEATEDLY with dose 2 and 3), so I think that it is even worse than it looks here
(that's what I get for doing this fast. This also doesn't take into account the one-week-
following-a-dose issue.
Larry Schonberger to Walt Orenstein, 6/23/99
As important as the Kaiser data may be, the way they are presented in the MMWR
is at least as important. Both deserve more attention. The current draft could be
misinterpreted by some as CDC's not fully appreciating the importance of using person-
time analyses and assessing intervals between vaccinations and onsets.
MW to John Livengood 6/24/99
Peggy Rennel's data. Using the NY State rate of 0.51 cases per 1,000 infant-years,
the expected number of cases within one week following vaccine (assuming all 10,054
vaccinated children got 3 doses) is 0.3, 95% CI 0.28, 0.31. With an observed of 3, this is
highly significant. For the 16,680 doses in NCK as of May 10, the expected number of
cases within one week is 0.16 (0.15, 0.17). If the period of observation is extended one
week, observed = 1, which is again significant.
To MW from John Livengood 6/24/99
I actually mentioned this to Ben yesterday. With the apparent misstatement of the
background rate and this, he was openly wondering what was up with this. She is not a
newby at this sort of thing. (Actually, I have the 0.16 written on a napkin when I was
trying to argue with Roger about the article and the inclusion of the NCK data.) I guess I
am extremely critical of the "inability" to extend the period for one more week (actually
one more day of rotavirus immunization) and wondering why this seems so coincidental.
Anyway, for now, we'll just chalk it up to us (with an assist from Larry) just being
smarter than everybody else! How will we deal with this in MMWR? I feel hesitant to
mention the statistical testing given that we believe it to be wrong.
I guess I am ok with the bending over backwards tone of the article, but we are getting
close to the line of discomfort for me. There is going to be something here and we
quickly are going to get clobbered.
To MW from Larry Schonberger 6/25/99
If the planned date of publication for this article is July 9th then you might want to
consider suggesting to the MMWR folks that they prepare a back-up article to reduce the
negative consequences should this article need to be held back. Conceivably, additional
case reports to VAERS and/or additional cases identified through active postlicensure
surveillance to N Calif Kaiser Permanente before July 9th could lead to a decision to
quickly call an executive committee together of the ACIP to reconsider the earlier
decision to continue to recommend the routine use of RRV-TV for vaccination of healthy
infants; Walt Orenstein might want the option to postpone the publication for a week so
the ACIP can reconsider the latest analyses of any new data without causing undue
hardship at the MMWR. Larry S.
from MW to Frank Destafano, 6/26/99
Charlie just told us that a trial in China in the 1980s was stopped because of
occurrence of intussusception in rotavirus vaccine recipients (a different vaccine than the
currently licensed product).
Melinda Wharton (MW) 6/28/99 to John Livingood, Trudy Murphy, Rebecca Prevots, Mehran
Massoudi
In order to include the case reported to VAERS, they are having to look at
admission diagnoses because for some reason intussusception was not included among
the DISCHARGE diagnoses. When they include admission diagnoses there are a bunch
of other potential cases (I think he [Steve Black] said 11) and charts will have to be
reviewed. So much for completeness of ascertainment by review of discharge data, and
for VAERS being "biased" by the active surveillance project -- the influence seems to be
going in the other direction.
MW 6/30/99 to Trudy Murphy, Rebecca Prevots, Lynn Zanardi, Charles Vitek, John Livengood,
Mehran Massoudi, Rick Nelson
...people thought that infants transferred to surgical services might very well receive
discharge diagnoses of "s/p bowel resection"....NCK found ANOTHER nonascertained
case by reviewing the barium enemas done on infants, that didn't show up from EITHER
discharge or admission codes.
MW to Rick Nelson 7/2/99
The MMWR is still scheduled for publication on July 9 (but that could change).
Seven states are included in the table in the article as having cases reported to VAERS
(MO, NY, PA, NC, CA, CO, KS). The MMWR editor reminded me this afternoon that
the state epidemiologists in these states needed a heads up that this was coming out.
There is some overlap with the study states, so several of them already knew all about it
(but not necessarily that there was a reported case from their state, but some hadn't heard
anything).
Trudy Murphy to MW 7/2/99: need to check with parents because of multiplicity of vaccine
providers, including some who record only vaccines given at their own facility.
MW to Larry Schonberger 7/7/99
Thanks again for all the help in getting the article in good shape. I'm concerned,
though that Rennels et al. are going to be blindsided by the data included, which would
lead an astute reader to conclude than an inappropriate analysis was done in their
published paper. If I were them, I'd rather point it out myself before someone writes a
letter to the editor of PIDJ that I have to respond to....
MW to John Livengood 7/7/99
As you know, there are currently more than 10 states involved in the ongoing
investigation, and many people know about the rotavirus issue. We have tried to keep the
lid on this, in terms of public announcements, at least until the MMWR was published,
but with the delay that may no longer be possible. What are we supposed to say to states
regarding media statements and other issues?
John Livengood to Claire Broome and Dixie Snider 7/7/99
Now that it seems we have a joint mercury statement to publish in MMWR, what is
your final determination on the rotavirus MMWR. I couldn't hear clearly yesterday on the
call from the Thurm meeting, but he was concerned about the confluence of issues. On
the other hand, we need to get the rotavirus out soon to be responsive to what we
promised ACIP, and John Ward believes that it would be a better communications
strategy to run both and not prolong the "bad news."
from Susan Chu to Rebecca Prevots, Trudy Murphy, Carol Goettl and MW, 7/13/99
In final form, Marjorie can forward [protocol] to Tom with her justification that this
investigation is not research....
Trudy Murphy to MW 7/14/99 re: Jose C comments on MMWR:
The first paragraph, second sentence indicates that the recommendation for use of
RRTV-TV was from the ACIP and the AAP. Was it not also from the AAFP?...
The last paragraph recommends postponing use of the vaccine, and then that providers
should warn parents of the signs and symptoms of intussusception if the vaccine is used.
This is a permissive recommendation, and in Jose's view, substantially different from the
AAP recommendation to completely suspend vaccination.
Barry Sirotkin to MW 7/16/99 re: does anybody die of intussusception? There were 19 deaths in
1994, 29 in 1995, 21 in 1996.
Dixie Snider to Mark Messonnier 7/19/99
The scenario here is that we may wind up in a position of proving a causal
association between rotavirus vaccine and intussusception and yet the event may be very
rare. An economic analysis may favor continuing the use of the vaccine. If, based on such
findings, CDC and ACIP were to decide to recommend resuming vaccination, we could
be in for a firestorm since we are, in some people's view, vaccine "hawks." So, the
question of involving the CDC ethics subcommittee and economic analysts who are not
associated with CDC or vaccines could help give greater credibility to the result,
whatever it is. Is that basically what your concerns are, Walt?
MW to Larry Schonberger, 7/21/99
I have been assuming that it would be a short interval (missed diagnosis, shock from
dead bowel, etc.) but I could be wrong about that. However, if a child ended up being
hospitalized for a month with peritonitis post-operatively (for example) it would lower
the case fatality ratio in the younger age group -- so the estimates will be conservative.
FYI, we did this already using the VSD rates and as expected the case fatality ratios are
higher for infants < 6 months old than 6-11 months old. John did a back-of-the-envelope
calculation that suggests if the vaccine does this 1 in 10,000 doses (our current guess)
then the deaths due to intussusception will be of the same magnitude as the deaths due to
rotavirus vaccine -- which is not a good thing, if true.
MW to Walt Orenstein, John Livengood, 7/22/99
A fatal case was reported this morning; a 5-month-old previously healthy child
received rotavirus, DTaP, IPV and Hib/Hep B vaccines on February 18, 1999. On
February 23, the child was seen in the emergency room and diagnosed with
gastroenteritis. The following day she developed bloody diarrhea and was seen in a clinic.
She died while awaiting helicopter transport to a pediatric intensive care unit. At autopsy,
the child was found to have an 8 cm intussusception with necrosis and septic shock. The
case was reported from Minnesota, but was not one that had previously been identified by
the health department.