Jane Orient, M.D.
Association of American Physicians and Surgeons, Inc.
1601 N. Tucson Blvd., Suite 9
Tucson, Arizona 85716-3450

Dear Dr. Orient:

This letter is in response to your Freedom of Information Act (FOIA) request of October 4.

We received your check for $56.20. Enclosed are documents you requested.

Under authority of the Freedom of Information Act at 5 U.S.C. 552(b) (6) and the Department's implementing regulation at 45 CFR 5.67, names and other information, the disclosure of which would constitute a clearly unwarranted invasion of personal privacy, have been deleted from these documents.

Under authority of 5 U.S.C. 552(b) (2) of the Freedom of Information Act and 45 CFR 5.63 of the Department's implementing regulations, we are withholding internal information such as conference call bridge numbers. This information relates solely to the internal personnel rules and practices of this agency.

We are withholding predecisional internal communications. Release of this type of material would interfere with the agency's deliberative process. This decision is based upon the Freedom of Information Act at 5 U.S.C. 552(b) (5) and the Department's implementing regulation at 45 CFR 5.66(a).

You have the right to appeal this decision to deny you full access to agency records. Should you wish to do so, send your appeal, within 30 days from the date you receive this letter, to the Deputy Assistant Secretary for Public Affairs (Media), U.S. Department of Health and Human Services, Room 17A-46, 5600 Fishers Lane, Rockville, Maryland 20857. Please mark both your appeal letter and envelope "FOIA Appeal."

We may have additional releasable records. If you are interested in receiving these, let our office know and we'll prepare an invoice as soon as we know what is available.

Portions of documents generated by either the Food and Drug Administration (FDA) or the National Institutes of Health (NIH) were deleted from these records. You should contact those agencies directly for any further information.

Sincerely yours,

Lynn Armstrong
Office of Communication
(404) 639-7270


From: Wharton, Melinda
Sent: Wednesday, June 30, 1999 3:15 PM
To: Murphy, Trudy; Prevots, Rebecca; Zanardi, Lynn; Vitek, Charles; Livengood, John; Massoudi, Mehran; Nelson, Rick
Cc: Reynolds, Barbara S.; Allen, Curtis
Subject: discussions at CSTE

Although I didn't have a chance to get any more states recruited (didn't see anyone from Indiana), we did discuss rotavirus vaccine issues over lunch yesterday, thanks to Sandy. Basically a good, constructive, supportive discussion. Folks from a number of states were there (MN, OH, OR, TN, NM, NJ are the ones I remember). We talked about everything from press response (lots of concern about when this hits the press) to how controls should be selected. They made the following suggestions:

1. There needs to be clarity an in process for ANY change in recommendations. Need to make clear that with publicity we expect additional cases, that the recommendation currently is that the vaccine still be used, but in the event that that changes, we will let people know immediately. There is concern about a lot of "local" decisions being made on the fly in lieu of a coordinated CDC response.

2. I told them that I was anticipating that the story would appear in the "trade tabloids" and maybe AAP in the July issues, and that these might be published before the MMWR story came out. Someone suggested we get those publication dates, and I've already talked to Barbara Reynolds about that -- she was going to try to have someone follow up on that.

3. There was a lot of conversation about "doing it fast" and "doing it right," and the tension between those competing priorities. Dave Fleming suggested that once the cases are ascertained from hospital discharge data that the FIRST step (before reviewing hospital charts, radiology reports, or anything else) should be obtaining vaccination histories (and if that means interviewing providers and/or parents, administering the complete questionnaire at the same time -- he was not advocating going back to a given source). There was general agreement that this was feasible and reasonable. These cases in vaccinated children would be tallied and could lead to public health action if warranted.

4. We need a procedure for handling "stimulated reports" in states participating in the study (and other states, for that matter). Once the MMWR article comes out, or press coverage that the state health dept is participating in the study, there will be providers or parents calling in "that happened to my child," but the child may have been the only child hospitalized at Hospital A for intussusception in the last 4 years, and the case would NOT have been ascertained. We didn't talk a lot about what that procedure should be, other than to say that the cases should be tabulated, but not included in the study. Approaches might be for the state file a VAERS report; for the state to encourage the person calling to file a VAERS report (not optimal); or to actually use the study questionnaire in case by chance the child's name DOES come up (so they don't have to call back and then the parent thinks the state is incompetent because they didn't ask all these questions three weeks ago when they reported the case the first time). This needs some thought and discussion.

5. Local health dept folks need some information on the benefits of rotavirus vaccine, and on the issue generally, because they will get asked questions.

6. An EIS officer said it was "awkward" that they knew about the study and their state immunization program folks DIDN'T -- which is when I realized that we hadn't pursued that channel simultaneously in either the EIP states or Epi-aid states (obviously in some states it's all the same folks, but in some it isn't). I don't know where we are with state contact at this point, but if we haven't followed up with the immunization programs in the states where we are trying to get things launched, I think we should. Rick, is that something you could do?

7. There was some discussion about control selection, much of which echoed discussions we've had here (i.e., need to age match, but don't want to match on much else, and birth certificate info is not going to be available at the state level in a timely way in all areas). Karen White from MN suggested using COUNTY level birth certificate matches (many of the advantages of using state data, but she thought this might be feasible in all states, since most of the delay is NOT in getting the information to the county, but getting it from the county to the state). Also would facilitate age-matching, since most counties would have enough births this would be easy, while many hospitals might not. Seems like a reasonable alternative to me but I have no idea whether all areas could do it. Probably some more discussion about selecting controls in order, and we will need very clear instructions on how EXACTLY it is to be done, once that decision is made.

8. I told them about the issue at NCK of admitting dx vs discharge dx, and people thought that infants transferred to surgical services might very well receive discharge diagnoses of "s/p bowel resection" or something (which is what I would guess happened with the NC K). There was interest in trying to do some validation in some places where it was feasible, and someone from MN called one of their pediatric hospitals right after the discussion to ask if they could check admitting dx vs discharge dx. This is an issue which deserves some more discussion (and it's actually even worse than we thought - NCK found ANOTHER non-ascertained case by reviewing the barium enemas done on infants, that didn't show up from EITHER discharge or admission codes, according to the msg Steve Black left today).

I also made a note to send out the updated background material. Does someone have a group code to send it out?


From: Wharton, Melinda
Sent: Thursday, June 10, 1999 11:48 AM
To: Livengood, John
Subject: expected rates

Expected rates for 300,000- 1,200,000 first doses. if it looks okay, please forward to Walt -.



Rate Lower 95% CI Upper 95% CI
expected rate among children <12 months of age 0.000389 0.000331 0.000456
Assuming number of first doses administered 300,000
Number of infants <12 months of age in U.S. 4,000,000
Expected number of cases of intussusception among infants <12 months of age (annual) 1,556 1,324 1,824
Proportion of infants receiving first dose 0.075
Expected number of cases of intussusception among infants receiving first dose 116.7 99.3 136.8
Expected number of cases of intussusception among infants within I week of first dose 2.2 1.9 2.6
Rate Lower 95% CI Upper 95% CI
expected rate among children <12 months of age 0.000389 0.000331 0.000456
Assuming number of first doses administered 600,000
Number of infants <12 months of age in U.S. 4,000,000
Expected number of cases of intussusception among infants <12 months of age (annual) 1,556 1,324 1,824
Proportion of infants receiving first dose 0.15
Expected number of cases of intussusception among infants receiving first dose 233.4 198.6 273.6
Expected number of cases of intussusception among infants within 1 week of first dose 4.5 3.8 5.3
Rate Lower 95% CI Upper 95% CI
expected rate among children <12 months of age 0.000389 0.000331 0.000456
Assuming number of first doses administered 900,000
Number of infants <12 months of age in U.S. 4,000,000
Expected number of cases of intussusception among infants <12 months of age (annual) 1,556 1,324 1,824
Proportion of infants receiving first dose 0.225
Expected number of cases of intussusception among infants receiving first dose 350.1 297.9 410.4
Expected number of cases of intussusception among infants within 1 week of first dose 6.7 5.7 7.9
Rate Lower 95% CI Upper 95% CI
expected rate among children 12 months of age 0.000389 0.000331 0.000456
Assuming number of first doses administered 1,200,000
Number of infants <12 months of age in U.S. 4,000,000
Expected number of cases of intussusception among infants <12 months ofage (annual) 1,556 1,324 1,824.
Proportion of infants receiving first dose 0.3
Expected number of cases of intussusception among infants receiving first dose 466.8 397.2 547.2
Expected number of cases of intussusception among infants within 1 week of first dose 9.0 7.6 10.5

From: Wharton, Melinda
Sent: Monday, June 14, 1999 5:22 PM
To: Kramarz, Piotr
Cc: Livengood, John
Subject: FW: Intussusception Questions that need answers
Importance: High

See the second set of questions about VAERS. Is this something that you all already have put together somewhere?


Original Message

From: Nowak, Glen
Sent: Monday, Junel4, 1999 1:56 PM
To: Livengood, John; Wharton, Melinda; Schwartz, Ben (NIP)
cc: Cordero, Jose; Landry, Martin; Nichols, Bill; Orenstein, Walt; Reynolds, Barbara S.; Thompson, Charlis J.
Subject: Intussusception Questions that need answers
Importance: High

Barbara Reynolds, Charlis Thompson and I have identified some of the key questions that the media are likely to raise. Though you might have already identified these, here's our suggestions:

I realize this is a fairly long list, with many difficult questions, but given these are the ones most likely to be asked, it's better to be prepared in advance. Let me know if you have questions or need additional information.

From: Wharton, Melinda
Sent: Thursday, June 17,1999 9:40 AM
To: Anderson, Larry
Cc: Livengood, John
Subject: RE: rotavirus vaccines

I tried to call you yesterday --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 is going to turn out. We'll see what happens this afternoon. Even if ACIP supports the current plan of action (continuing to use 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 just taking this one day at a time.


Original Message
From: Anderson, Larry
Sent: Wednesday, June 16, 1999 1:24 PM
To: Wharton, Melinda; Livengood, John
cc: 'Joseph Bresee'; Schonberger, Larry; Khabbaz, Rima
Subject: rotavirus vaccines

I am concerned about possible directions for communications from CDC and some less than enthusiastic support for rotavirus vaccines. I hope the luke warm support for rotavirus vaccines by some does not detract from appropriate interpretation of the data. Since I am about to leave for vacation, I wanted to put in my two cents worth. I will also be in close contact with Joe.

My read of the data is that:

1. the rate in the Kaiser study (only reasonable data) and VEERS data (pretty difficult to have much confidence in rates) both suggest no increase in intussuception with rotavirus vaccination.

2. given the small number of patients in Kaiser study and unknown numerator and denominator for VEERS data and, previous questions regarding intussuception with the vaccine, we can not be sure that rotavirus vaccination does not occasionally cause intussuception and must do further studies.

3. Given that the data to date is neutral or no increased risk, vaccination should proceed as before. It is reassuring 4 that intussuception is not life threatening illness (rarely requires surgery with present medical practice according the pediatricians among us). It would not be good to stop a vaccination program that can have substantial public health impact on the basis of data that raises a question but does not show a problem.



From: Wharton, Melinda
Sent: Wednesday, June 23, 1999 10:41 AM
To: Livengood, John

Sorry it took me so long to get this right -- had the wrong numbers of doses in the original calculations, but I think this is right. Bottom line: with these assumptions (equal rates of vaccine use early on to later, which I'm sure is wrong) and a baseline rate of 0.4 per 1,000, the expected number of cases among vaccines is 1.14, and within one week of any dose is 0.127. The intervals for the one case that has been reported to VAERS is 14 or 15 days, but the kid with the polyp (not reported to VAERS) was (I think) 8 days. The child that has been reported to VAERS but not included in the cases analyzed to date had a 3-4 day interval.

Mike said that they (NCK) are working on the person-time, and I clearly scared him this morning when I sent him an earlier (wrong) version of this. So they will be leaning on NCK to do it, as well, if only to show me wrong. I did tell him we couldn't include the NCK data without person-time, but if we got it we could put it in the body of the report.

Expected rates of intussusception among RRV-TV (recipients, Northern California Kaiser

Dose 1 9,860
Dose2 5,207
Dose3 1,613
Total 16,680

No. dose 1
Dec 1643.3 6 9,860.0
Jan 1643.3 5 8,216.7
Feb 1643.3 4 6,573.3
Mar 1643.3 3 4,930.0
Apr 1643.3 2 3,286.7
May 1643.3 1 1,643.3

Total infant-months 34,510.0
Total infant-years 2875.8
Expected rate among infants per infant-year 0.000389
Expected cases among vaccine recipients 1.12
Total doses of vaccine administered 16,680
Expected rate among infants per infant-week 0.0000076538
Expected number of cases within one week of any dose 0.127

From: Schonberger, Larry
Sent: Wednesday, June 23, 1999 3:40 PM
To: Orenstein, Walt
Cc: Livengood, John; Cordero, Jose; Khabbaz, Rirna; Wharton, Melinda; Bresee, Joseph
Subject: RE: Comments on the draft MMWR article about intussusception and rotavirus vaccine

Walt: Yesterday, Rima shared with me a copy of a draft MMWR article which contained a one sentence report of postlicensure surveillance data from the Northern California Kaiser. If I understand that data correctly, they, like the VAERS data, are not reassuring and I recommend that they not be presented as they are now in the current draft of the MMWR article.

First, let me update you on the background incidence of intussusception from New York based on information I asked Bob Holman to obtain for us. For the seven year period, 1991 through 1997, (note that we have added two more years of data) there were 972 hospitalizations of infants under 1 year of age coded for intussusception. Using census estimates of the population under 1 year of age in New York for each of the 7 years in the study, we can estimate the denominator as 99 million person weeks of risk, hence the estimated rate in infants in New York becomes 0.98 hospitalizations for intussusception per 100,000 person weeks of risk.

With regard to the Northern California Kaiser data, I will assume from the current draft MMWR article that Kaiser had, on average, a population of 27,860 children under 1 year of age that were followed for 23 weeks beginning December 1, 1998 and that 9,860 of these children received RRV-TV during this 23 week period. As I know you well understand, the timing of vaccinations is critical for calculating incidence rates. If the number of vaccinations administered was stable each week throughout the period, for example, then the incidence rate for vaccinees based on the 2 cases becomes 1.76 cases per 100,000 person weeks of risk. If the number of vaccinations administered increased over time such that, for example, the average vaccination was given only after 60% of the 23 week period had passed, instead of by the midpoint of the period, then the incidence rate for vaccinees based on the 2 cases becomes 2.2 cases per 100,000. Because the person week denominator is larger for the unvaccinated population, the background incidence rate for this population remains relatively stable with either of the assumptions above at about 1.1 cases per 100,000 person weeks.

When I spoke to Melinda Wharton this morning, she mentioned that if one extended the time period of the analysis of the Kaiser data by 1 week, the numerator for the vaccinees would include case number 12 of the MMWR Table. Such a case would, of course, increase the overall incidence rate among vaccinees and would add to the one week post vaccination incidence rate, the period of most concern based on cases reported through VAERS. Of course, the one week extension of the Kaiser followup study period could also change the numbers of people vaccinated and possibly the numbers of unvaccinated cases, but again, with what I would consi reasonable assumptions about how these numbers might change, the Kaiser study data are not reassuring to me. As important as the Kaiser data may be, the way they are presented in the MMWR is at least as important. Botheserve more attention. The current draft could be misinterpretted by some as CDC's not fully appreciating the importance of using person-time analyses and assessing intervals between vaccinations and onsets.

Let me suggest that you give me a call when you get a chance during your hectic schedule. My pager is 404 716- 2271. Thanks. Larry S.

---Onginal Message----
From: Schonberger, Larry
Sent: Thursday, June 17, 1999 8:10 PM
To: Orenstein, Walt
cc: Livengood, John; Cordero, Jose; Khabbaz, Rima; Schonberger, Larry
Subject: Comments on the rotavirus vaccine associated intussusception data and protocol

Walt: John and Rima both mentioned that you had wanted me to look at the key data on the association between rotavirus vaccine and intussusception and provide you with some feedback. I am not an expert on intussusception but clearly risk factors for this illness that might also influence likelihood of receipt of vaccine are factors that would be good to identify early. I will assume that there are no such overwhelming confounders, that the estimate of the number of first doses administered is 1.2 million (that is; 1.8 million vaccinations distributed, 1.5 million administered including 0.3 million second or third doses, and that there has been a decision that vaccinations will continue at least until John Livengood et al can obtain epidemiologic data to better define whether there is an etiologic association between vaccination and intussusception and if so the magnitude of any risk.

Because I have more raw data from the New York State study (all hospitalizations between 1991 and 1995), I will use these data as my estimate of the background rates of intussusception for the entire country. This study yields higher background rates than the ones I believe you have been using but I do not think this will significantly alter our main conclusions.

The estimated background incidence rate in New York in terms of cases per 100,000 person weeks of risk for those under 1 year of age is approximately 0.97. From the New York data, the background incidence rate for the specific age group 2 months through 6 months of age is estimated to be about 21 % higher than this overall rate, or 1.18 cases per 100,000 person weeks of risk. The 2 months through 6 months of age period represents the most likely period when an infant would receive rotavirus vaccinations.

Using the above relatively high background estimates, nationally the 1.2 million first doses should produce an estimated 11 (if the 0.97 rate above is used) to 14 cases (if the 1.18 rate above is used) of intussusception in children under 1 year of age in the 1 week period after a first dose of rotavirus vaccine; I understand that the VAERS system has identified at least 9 such cases whose diagnoses have been confirmed. If the sensitivity of VAERS reporting were 64% to 77% these ascertained cases would not suggest that there existed a true excess of cases in this critical period of concern. If this sensitivity of reporting, however, were 50% and there were in fact 18 cases of intussusception (9 more not yet identified), then we would estimate that nationally there were between about 3 to 5 extra cases of intussusception per million first doses of vaccine. Similarly, if the sensitivity of VAERS were 10%, then we would estimate that there were between 63 and 65 extra cases per million first doses of vaccine. These analyses are not reassuring.

Despite the signal from VAERS about a potential problem with first doses, in the follow-up study that is currently being planned, I agree with broadening our concern to any dose of rotavirus vaccine. I say this despite my suspicion that for a variety of reasons we might expect that if there were a risk of intussusception after receipt of rotavirus vaccine then that risk would be more apparent after first doses than after subsequent doses. The risk period of key concern Is appropriately the one week period after receipt of any of the doses of vaccine and, as may be planned for the follow-up study, there will and should be analyses of data by risk windows slightly longer than 1 week as well.

Do I agree with the case-control methodology? Yes, because a) we need cases from a huge population, b) there are problems with pinning down appropriate denominators and C) there likely is wide variability in vaccine coverage in different geographic areas. However, I would anticipate that completing such a case-control study would take considerable time and effort before it would likely provide us with data about whether there is really an increased vaccine-related risk of intussusception, and if so, some information about the magnitude of this risk.

Since children are still being vaccinated, I would suggest that we do the case-control study but break one of the usual rules of such studies and focus initially on just ascertaining cases from throughout the country and determining their vaccination status. We can begin to identify the controls and prepare for their eventual inclusion into the case- control study, but initially give the case ascertainment almost exclusive priority. The justification for this would be to hasten determination of whether there is likely a risk and particularly to hasten detection of a relatively large risk should it exist. For example, you might arbitrarily set in advance some arbitrary number of new cases in the first week after a rotavirus vaccine that, if ascertained in this initial case ascertainment effort, would lead you to temporarily halt the vaccination program or to at least put out more information about the fact that some risk exists.

One negative consequence of initially working up cases without simultaneously working up the controls is our potentially creating a systematic bias in the histories of vaccination; that is, controls will potentially be interviewed about their vaccination history months after their matched case will have been. Given that we would plan to confirm the histories of vaccinations by checking physician records, this systematic bias is less worrisome to me and probably does not outweigh the benefits of our focusing on quickly ascertaining as many cases of intussusception as possible. This latter ascertainment could also include tapping into data of health maintenance organizations or other institutions that might have local unvaccinated and vaccinated denominators that would facilitate interpretations of risk.

In the cooperative, NIP/DVRD study of GBS in 1990-91, DVRD actually took primary responsibility for the HMO aspect of case ascertainment and provided an additional EIS Officer to help out with other case ascertainment efforts. A similar type of cooperative arrangement might be considered for this investigation too, although I have not really discussed this with Dr. Larry Anderson. Since Dr. Anderson's group, I suspect, would be interested in the possiblity that even should rotavirus vaccine trigger some instances of intussusception, this might be similarly true for wild rotavirus disease as well. Longer term follow-up of large numbers of vaccine-recipients might show, for example, some long term protective effect of the vaccine against developing intussusception in the period during which recipients would be relatively protected from the wild rotavirus disease. The HMOs might give Larry Anderson's group some opportunity to look for evidence of this hypothetical possibility though longer term follow- up of vaccine recipients.

The tremendous work involved in conducting the case-control study argues for not wasting too much effort on cases and controls who by their date of birth would be at no or extremely low risk of receiving rotavirus vaccine. The current protocol, for example, does not exclude children who were born between December 1997 and April 1998 even though these children will have been beyond the recommended age for vaccination during the likely period when vaccine would have been available for the infants who could become cases in the follow-up study period, November 1998 -June 1999. Although it is true not all physicians will necessarily follow ACIP recommendations on the age group to vaccinate and possibly begin to vaccinate infants 7 months old or older, I suggest excluding these children from the case-control study.

In summary, the current data are not reassuring to me. A rotavirus vaccine-related risk of intussusception on the order of magnitude discussed above would not be surprising and should be ruled out as quickly as reasonably can be done. Although the uncertainties about denominators argues for the case-control methodology, the need for a more rapid assessment argues for initially focusing almost exclusively on case ascertainment nationally and taking advantage of any subgroup denominators that may exist, as in large HMOs. A decision should be made early about what number of vaccine-associated intussusception cases would lead you to recommend temporarily halting rotavirus vaccinations or to revise information provided to parents deciding whether to vaccinate their infants. Get Larry Anderson's group involved in some of the HMO case ascertainment aspects of the study. Drop birth cohorts from the case-control study who would have been too old to have likely received rotavirus vaccine by the time the vaccine became reasonably available.

Hope these comments are helpful.
Larry S.

From: Wharton, Melinda
Sent: Friday, June 25, 1999 9:15AM
To: Livengood, John
Subject: RE: MMWR

I got the changes and they're good (and I think readable). Only other thing I was thinking about was explicitly stating that the expected number of cases within one week of vaccine in the prelicensure trials was <1. That can be calculated from the data in the report, and it will be an obvious omission if it is not there.

"However, three of five cases among vaccinated children occurred with onsets of 6-7 days of receipt of rotavirus vaccine; based on observed background rates of intussusception, <1 cases was expected during that period."

The other thing I'm getting an ulcer about is how this whole thing was presented to AClP. Would the outcome have been different if we had pointed out that the cases within one week were 10-fold higher than expected, and that the NCK data (including the next week) also were expected, even if n=1? Should we tell them now, before this goes to print?

Steve Black called me last night; they will be extending the period of analysis to May 17, so it will include the 3rd case. Anyway, I'll make the edits and get it off to Dr. Koplan.


----Original Message
From: Livengood, John
Sent: Friday, June 25,1999 8:31 AM
To: Wharton, Melinda
Subject: MMWR

I arn faxing the 2 pages with final cornrnents. Please make these last changes (for this round) and send to Jeff, cc Martha, Dixie and Walt Jose Roger and me. Note should say this is the draft for his comments. preliminary review by NClD/FDA, but not final approval. Will add NCK data if it becomes available early next week. ask for comments by Monday (noon?) and inquire how and when to sent to Mr. THurm. Reminder that this will need to go to MMWR by Wed to be published 7/9 due to shortened holiday week.


From: Glass, Roger
Sent: Friday, September 03, 1999 10:47 AM
To: Adams, Melissa; Speers, Marjorie
Cc: Shefer, Abigail; Rodewald, Lance; Glass, Roger; Livengood, John; Anderson, Larry
Subject: RE: \HSREB-N IP\Rotavirus-ethics\rotavirus_ethics.wpd

Melissa and Abigail,

I read over the Ethics discussion and was pleased by the way you have approached the alternatives but not sure that the data presented is in fact correct. Below are some issues for consideration:

1. Are your numerical estimates of IS cases related to RV plausible? On page 5, you suggest that 6000 children per year would develop intussusception over baseline with continued use of the vaccine and 37 deaths. The excess risk of IS noted in the studies quoted above refer to excess risk in the week or 3 weeks after immunization-- not to an excess for an entire year or a child's first years of life which would be much less. Consequently, in the Kaiser study, 9000 vaccinated children had 2 episodes in the 3 weeks after vaccination--that might be extapolated to about 2/10,000 or 200/million or 800 total for the US birth cohort of 4 million. So the estimate of 6000 seems very high. I might ask how that figure was obtained.

2. Knowledge of intussusception might radically change outcome. If physicians were informed to think about IS as a complicaton, the risks of surgery or death might be significantly diminished. In many settings, few children end up in surgery and the risk of a fatality is low ( on the order of 1/500) or several deaths from excess vaccine related cases. Clearly, knowledge of more precise figures can alter the direction of the assessment.

3. In India, 100,000 children will die of RV each year (about 1 in 300 children born) so a vaccine that has a low risk of IS (say 2/10,000) and a smaller risk of fatal IS would still be a life-saver. At issue is replacing one fatal risk which we induce by vaccine for a greater risk that we prevent. Can we ethically deny a vaccine from these children given that their risk of fatal rotavirus is so great? Should we test these vaccines further in developing countries given the risk that we will soon learn for American kids? Knowing that intussusception is a complication that could require surgery for repair or be fatal, could we test a vaccine in India in a setting where surgery is available (eg. catchment area of a hospital) and in a setting where health workers visit, educate and bring in vaccinees who present on day 3-10 with symptoms of obstruction?

4. On page 2, you state a number of complications associated with RV but don't mention intussusception. In 4 studies at least, 8-38% of IS cases have RV detected in stool or antibodies in sera. So RV vaccines may well protect against this outcome. Causality still has to be addressed and this will be difficult to do but not impossible. The bottom line is that the vaccine may well diminish this adverse event.

All vaccine testing of live oral vaccines has stopped because of fear of litigation and risks of complications. The current vaccines could well be life-savers in developing countries and during the 6-8 years it may take to develop the next generation of rotavirus vaccines, 4-6 million children will die of rotavirus diarrhea--some 2000 per day. Any ethical review should address the potential considerations for considering continued use of the vaccine in developing countries.

Happy to discuss this further but I think that the piece as presented should highlight these developing country issues and put more appropriate estimates on the extent of the IS problem in the US or leave these blank until we have more accurate information.

Look forward to opening a dialogue on these ethical considerations. Hope these comments help.

Roger I. Glass, M.D., Ph.D.
Chief, Viral Gastroenteritis Section
Division of Viral & Rickettsial Diseases
MailStop G04
Centers for Disease Control & Prevention
1600 Clifton Rd., NE
Atlanta, GA 30333 USA
+1 1-404-639-3577
+1-404-639-3645 (FAX)

Original Message
From: Adams, Melissa
Sent: Wednesday, September 01,1999 5:01 PM
To: Speers, Marjorie
Cc: Shefer, Abigail; Glass, Roger; Rodewald, Lance
Subject: RE: \HSREB-N I P\Rotavirus-ethics\rotavirus_ethics .wpd

Sorry for the delayed response, Marjorie. I wanted to check w/folks here to see how much time they needed for review before sending it to your shop. I thought that Abby would be sending you something today, but Roger Glass stopped by this afternoon and made a compelling arguement that the document should include something about the potential impact on developing countries of a decision to not recommend the vaccine. [The risk of death from diarrhea is much different in developing countries than in the U.S. A decision to not recommend the vaccine would very likely halt the development of competing products by other manufacturers. Without competition, a cheap vaccine will not become available and the potential spill-over benefit to children in developing countries will not be achieved.] I've asked Abby to talk w/Roger and incorporate his comments. We plan to have something to you tomorrow.

Original Message
From: Speers, Marjorie
Sent: Monday, August 30,1999 11:20 AM
To: Adams, Melissa
Subject: RE: \HSREB-N I P\Rotavirus-eth ics\rotavirus_ethics .wpd

Any sense on when a draft will be available?

Marjorie A. Speers, Ph.D.
Deputy Associate Director for Science
1600 Clifton Rd, NE MS D50
Atlanta, GA 30333
404 639-7341 (fax)

Original Message
From: Adams, Melissa
Sent: Monday, August 30, 1999 11:13 AM
To: Speers, Marjorie
Subject: RE: \HS REB-N I P\Rotavirus-ethics\rotavirus_ethics .wpd

Thanks for your helpful comments. Work on the document continues apace here. Stay tuned!

Original Message
From: Speers, Marjorie
Sent: Monday, August 23, 1999 10:39 PM
To: Adams, Melissa
Cc: Rodewald, Lance; Shefer, Abigail; Bernier, Roger; Snider, Dixie
Subject: RE: \HSREB-N I P\Rotavirus-ethics\rotavirus_ethics .wpd

I looked over the outline. The outline structure looks well thought out and to cover the major issues. I am more likely to be helpful once you have a draft of text. Here are some ideas for you to consider:

1. The ethical principles you have chosen are those used to guide research. They are focussed on the individual because the ethicists, at the time, were thinking about biomedical research and psychological research, two areas where the emphasis is on the individual. You may want to consider adding another principle which guides public health - balancing societal good with individual good. Although this principle may not be as relevant to the rotavirus vaccine, I think it should be discussed because we view immunization programs as public health programs.

2. Following on the former, you should discuss herd immunity or that lack of it.

3. When discussing the ethical principles, I would consider them from the viewpoints of children, parents, and providers.

4. When the ethicists were debating ethical prinicples over 25 years ago, there was a lively debate about including children in research. The basic premise is that children cannot give consent because they lack the mental decision making capacity to do so. Thus, we get "assent", general agreement, from children and "permission" from their parents. During the debate 25 years ago, one theologian argued that children should never be included in research because they cannot consent. Another theologian argued that parents could consent on behalf of their children because parents would act in the best interests of their children. A compromise was reached. Children could be involved in only minimal risk research or research that held the prospect of direct benefit to the child. Riskier research should not undertaken except under special circumstances. You may recall that Tom Murray mentioned once during the call that parents are acting on behalf on their children. He was speaking from some of the work he has done on this topic. But what I think you may want to consider is whether there is a conflict of interest for parents to give permission for their child to receive the vaccine when a major benefit of the vaccine is to the parent (in terms of reducing lost work time). This, to me, is one of the interesting, hard ethical issues with this vaccine. It may come up when you discuss autonomy and whether having parents be informed and consent obviates the potential harmful effects.

5. There hasn't been much discussion about the balancing of the 3 or 4 ethical principles; however it seems that autonomy is held a bit higher than the rest. Further, Western thinking seems to suggest that part of what makes an activity ethical is having the people involved appropriately informed and make a decision to participate. During the AzT trial debate, Peter Lurie argued that informed consent can never make an unethical activity ethical. One approach would be to discuss whether the rotavirus vaccination program in general is ethical and then to suggests that potential harms could be minimized through informed consent.

6. Consider comparing the rotavirus vaccine to other immunizations as appropriate. I found the comparison to OPV/IPV and pertussis helpful.

Original Message
From: Adams, Melissa
Sent: Thursday, August 19, 1999 4:20 PM
To: Speers, Marjorie
Cc: Rodewald, Lance; Shefer, Abigail; Bernier, Roger
Subject: L :\HS RE B-NI P\Rotavirus-ethics\rotavirus_ethics .wpd

Marjorie -- attached is proposed outline for the rotavirus ethics paper. We would value your comments. Abby will return on Monday and begin to develop the manuscript.

-- Melissa

From: Wharton, Melinda
Sent: Wednesday, July 14,1999 3:42 PM
To: Livengood, John; Orenstein, Walt; Cordero, Jose
Subject: FW: Possible Association of Intussusception with Rotavirus Vaccination

Original Message
From: Murphy, Trudy
Sent: Wednesday, July 14, 1999 3:01 PM
To: Livengood, John; Wharton, Melinda; Yacovone, Edward; Bresee, Joseph;
Massoudi, Mehran; Nelson, Rick; Prevots, Rebecca; Zanardi, Lynn Subject: FW: Possible Association of Intussusception with Rotavirus Vaccination


Original Message
From: AAP Office of the Executive Director [mailto:EXECOFFICE@AAP.ORG]
Sent: Wednesday, July 14,1999 1:55 PM
Subject: Possible Association of Intussusception with Rotavirus

Possible Association of Intussusception with Rotavirus Vaccination (Committee on Infectious Diseases)

New information indicates there may be an increased risk of intussusception during the first few weeks after receipt of the licensed tetravalent rotavirus vaccine (RRV-TV), RotaShield. Currently available data are very limited and are based on passive reporting to the Vaccine Adverse Event Reporting System (VAERS), a post-licensure study of adverse events at Northern California Kaiser, and active case finding in two states. Results thus should be considered preliminary. To date, the Centers for Disease Control and Prevention (CDC) has received reports of 23 cases of intussusception following receipt of doses 1, 2, or 3 of RRV-TV. The number of children who have received RRV-TV is unknown, however, the observed rate of intussusception among vaccine recipients during the first 3 weeks after immunization appears to be greater than expected, with the highest rate during the first week following vaccination. These initial data suggest that intussusception occurs at a younger age in vaccine recipients than in unvaccinated children.

The results of an ongoing case-control study of intussusception following rotavirus vaccination, conducted by the CDC, are anticipated to be available in a few months. At that time, a re-evaluation of the health risks and costs of rotavirus vaccination versus the health risks and costs of natural rotavirus infection will be performed. Since the seasonal risk of natural rotavirus infection in the United States will be very low during the next few months, the AAP is making the following interim recommendations:

1. Clinicians temporarily should suspend administration of rotavirus vaccine to unimmunized and partially immunized children, pending collection and evaluation of additional information.

2. Parents or guardians of children who have received RRV-TV within a period of approximately 3 weeks should be advised to promptly contact their physician if signs or symptoms compatible with intussusception develop.

3. All cases of intussusception which occur following administration of RRV-TV should be reported to VAERS (800-822-7967; www.fda.gov/cber/vaers/report.htm).

From: Wharton, Melinda
Sent: Thursday, July 15, 1999 5:01 PM
To: Schmitt, Tom
Subject: FW: AP Story on rotavirus

From: Thompson, Charlis J.
Sent: Thursday, July 15, 1999 4:46 PM
To: Orenstein, Walt,; Wharton, Melinda
Cc: Nowak, Glen; Allen, Curtis; Hibbs, Beth; Livengood, John; Reynolds, Barbara S.
Subject: FW: AP Story on rotavirus

CDC recommends suspending use of diarrhea vaccine
4:39 p.m. ET (2040 GMT) July 15, 1999

By Patricia J. Mays, Associated Press

ATLANTA (AP) - The government is recommending that doctors temporarily stop giving children a diarrhea vaccine because at least 20 infants have developed a bowel obstruction.

The Centers for Disease Control and Prevention said Thursday that RotaShield, the vaccine against rotavirus, has not been conclusively linked to bowel obstruction, but early studies show it may increase the risk.

The obstruction occurs when one part of the bowel becomes enfolded within another. Symptoms include vomiting, bloody stools and abdominal pain. Surgery is often needed to clear the blockage if it isn't caught early.

The CDC recommended suspending use of the vaccine in all children - including those who have begun the three- dose series - until November so health officials can conclude additional studies. Following the CDC's move, the American Academy of Pediatrics issued the same recommendation.

It's enough to make us worry," said Dr. Melinda Wharton, head of the CDC's child vaccine preventable diseases branch. "We're fortunate that rotavirus is a seasonal disease so we can postpone vaccination for a few months."

Rotavirus, an intestinal infection, is the leading cause of severe diarrhea in children under 5 in the United States. Each year, an estimated 3 million American children are sickened, 50,000 are hospitalized, and 20 die. The disease primarily occurs in winter and spring.

The vaccine's manufacturer, Philadelphia-based Wyeth-Ayerst Laboratories, said 1 million infants have been vaccinated with RotaShield since it received government approval in August. The three-dose series is given at ages 2 months, 4 months, and 6 months.

Before its approval by the Food and Drug Administration, five cases of the bowel obstruction were reported for every 10,000 vaccine recipients.

As a result, the bowel obstruction was listed as a potential adverse reaction on the package insert. The CDC said it is still investigating, but based on the early findings so far, it appears that the bowel obstruction rate may be higher than expected.