P R O C E E D I N G S (8:05 am)

Agenda Item: Welcome, Statement of Conflict of Interest, Acknowledge of New Members, Announcements

DR. FREAS: Members of the committee, invited guests, and public participants I would like to welcome you to this, our 82nd Blood Products Advisory Committee meeting. I am Bill Freas. I will be the acting Executive Secretary for today.

This meeting will be open to the public all today and half a day tomorrow. The meeting tomorrow, according to The Federal Register notice will go into closed session for discussion of personnel issues.

Before I go around and introduce the members seated at the head table, I would like to announce that we have five new members on the committee. They are: Dr. Matthew Kuehnert; Dr. George Schreiber; Dr. Keith Quirolo; Dr. Donna Whittaker; and Dr. Catherine Manno.

Now, I will go around the table, starting on the right-hand side of the table, and introduce all the members currently seated at the table. In the first chair we have Dr. Matthew Kuehnert, Assistant Director for Blood Safety, Division of Viral and Rickettsial Diseases, Centers for Disease Control and Prevention.

In the next chair we have Dr. Liana Harvath, Deputy Director, Division of Blood Diseases and Resources, NIH.

Next, we have Dr. Harvey Klein, Chief, Department of Transfusion Medicine, NIH.

In the next chair we have Dr. George Schreiber, Associate Director for Health Studies, Westat, Rockville, Maryland.

Next, we have Dr. Judy Lew, Assistant Professor of Pediatrics, University of Florida.

Next, we have Dr. Keith Quirolo, pediatrician, Department of Hematology, Children's Hospital and Research Center, Oakland, California.

Next is our chairman of this advisory committee, Dr. James Allen, President and CEO, American Social Health Association, Research Triangle Park, North Carolina.

Next we have Ms. Judith Baker. Ms. Baker is acting as our consumer representative for today. She is Regional Coordinator, Federal Hemophilia Treatment Centers, Region IX, Los Angeles, California.

Around the corner of the table we have Dr. Leonard Hudson, Professor of Medicine, University of Washington, Harborview Medical Center, Seattle, Washington.

Next we have Dr. Catherine Manno, Professor of Pediatrics, The Children's Hospital of Philadelphia, University of Pennsylvania, School of Medicine.

Next, we have Dr. Donna Whittaker, Director, Robertson Blood Center, Ft. Hood, Texas.

Next, we have Dr. Susan Leitman. Dr. Leitman is joining us today as a non-voting consultant. She is Chief, Blood Services Section, Department of Transfusion Medicine, NIH.

At the end of table we have our acting non-voting industry representative, Dr. Louis Katz. He is Executive Vice President of Medical Affairs, Mississippi Valley Regional Blood Center, Davenport, Iowa.

There are two committee members that could not be with us today. They are Dr. Kenneth Davis and Dr. Suman Laal. The others will be joining us shortly.

I would now like to read into the public record the conflict of interest statement for this meeting. The following announcement addresses conflict of interest issues associated with this meeting for the Blood Products Advisory Committee on March 17-18, 2005.

Pursuant to the authority granted under the committee charter, the director of FDA's Center for Biologics Evaluation and Research has appointed the following individuals as temporary voting members for the committee discussions: Ms. Judith Baker, and Dr. Liana Harvath. In addition, the Association Commissioner for External Relations, Food and Drug Administration has appointed Dr. Leonard Hudson as a temporary voting member.

Based on the agenda, FDA has determined there are no specific products being considered for approval at this meeting. Committee participants were screened for their financial interests to determine if any financial interests of conflict existed. The agency reviewed the agenda and all relative financial interests reported by the meeting participants. The Food and Drug Administration prepared general matters waivers for participants who required a waive under 18 US Code 208. Waivers were granted to: Dr. James Allen; Dr. Donna DiMichele; Dr. Catherine Manno.

Because of the general topics impact so many entities, it is not prudent to recite all potential conflicts of interest as they apply to each member. FDA acknowledges that there may be potential conflicts of interest, but because of the general nature of the discussion before the committee, these potential conflicts are mitigated.

Dr. Louis Katz will be participating as a non-voting industry representative on behalf of regulated industry for this meeting. Dr. Katz' appoint is not subject to 18 US Code 208. He is employed by the Mississippi Valley Blood Center, and thus has a financial interest in his employer. Dr. Katz reported that he is an associate investigator on the abbreviated donor study questionnaire. He also received speaker fees, and was a scientific advisor for firms that could be affected by today's committee discussions.

With regards to FDA's invited guest speakers, the agency has determined that these services are essential. For discussion of Topic 1 related to the safety of albumin, the following disclosures will assist the public to objectively evaluate presentations and/or comments made by the participants.

Dr. Paul Hebert is Chair, Transfusion and Critical Care Research, Ottowa Health Research Institute, University of Ottawa. Dr. Simon Finfer is Senior Staff Specialist in Intensive Care, Royal North Shore Hospital, Sydney; Clinical Associate Professor, Northern Clinical School, University of Sydney, and ANZICS Clinical Trials Group Executive.

For the discussions on Topic 2 related to the review of standards for plasma products for transfusion, Dr. Irma Szymanski is Professor of Pathology Emerita, University of Massachusetts Medical School, Worcester, Massachusetts.

FDA participants are aware of the need to exclude themselves from discussions involving specific products or firms for which they have not been screened for conflicts of interest. Their exclusion will be noted for the record.

With respect to all other meeting participants we ask in the interest of fairness that you state your name, affiliation, and the address any current or previous financial involvement with any firm whose products you wish to comment upon. Waivers are available under request through the Freedom of Information Act.

That ends the reading of the conflict of interest statement.

Before I turn the meeting over to the chair, I would like to ask if you have a cell phone or a pager or something, if you would put it in the silent mode. We would really appreciate it.

Dr. Allen, I turn the meeting over to you.

DR. ALLEN: Thank you, Dr. Freas.

And thank you for your very clear English statement on turning cell phones off or putting them silent. I get very annoyed with the airlines that tell you put it in the off position. I don't know exactly what that is.

Perhaps some of you attended congressional hearings yesterday, or listened to the news summary of Dr. Lester Crawford's interviews as the Commissioner for Food and Drug. I think that discussion certainly highlights the interaction between FDA as a regulatory body, the regulated community, manufacturers, blood collectors and others, the scientific community of practicing physicians and other who use products, and certainly the interests of the patients. This committee is charged to assist the FDA in its deliberations. And certainly listening to the discussion, I got a renewed sense of awe for the commitment that we have.

I do want to remind all of the speakers, including those who will be speaking during the open session, that we have a lot of requests. A very important part of the deliberations of the committee is an opportunity for discussion among ourselves, and consideration in full of the questions that are being raised before us. So, I would please remind everyone to keep within posted time limits. Make your points briefly, succinctly, and allow the committee then to discuss things.

I would also like to acknowledge the work of Dr. Bill Freas, Dr. Martin Riva, and others on the FDA staff who have picked up from Dr. Linda Smallwood after her retirement. I know she is significantly missed by the committee, and I greatly appreciate the work of everyone who has helped pick up after her retirement.

At this point I would like to turn the microphone over to Dr. Jay Epstein to make an announcement.

DR. EPSTEIN: Thank you very much, Jim.

Well, I have just one announcement, and it's a bit of a bittersweet task. That's to make know that Dr. Linda Smallwood retired from the federal service at the end of December, and in the same stroke concluded 17 years of service as the executive secretary of the Blood Products Advisory Committee.

In recognition of that singular contribution to the public health, it is my pleasure to award her a plaque from the Center for Biologics Evaluation and Research, which is an award of appreciation to Dr. Smallwood for exemplary service as the Executive Secretary to the Blood Products Advisory Committee from 1987 to 2004.

So, Linda, if you would like to come up, I would be pleased to present this.

[Applause.]

DR. FREAS: While they are taking that picture, I would just like to say for the new committee members who do not know Linda Smallwood, she really is to BPAC what Cal Ripken is to the Baltimore Orioles.

DR. EPSTEIN: And I just want to add my thanks to the new people who are joining the committee, as well as to those who are seated as temporary voting members. We know that it requires a special effort on your part to come and to participate, and to read through the voluminous materials that we provide to you. And just to say on behalf of the Center for Biologics Evaluation and Research how much we appreciate that effort, and how valuable the contribution is to our work in maintaining the safety, purity, and potency of blood and blood products.

Thank you.

DR. ALLEN: Thank you, Jay.

Dr. Smallwood, thank you for coming back. We miss you, and congratulations, and good luck in your retirement.

We'll move into the committee update session. First, will be Dr. Jerry Holmberg, Executive Secretary of the Advisory Committee on Blood Safety and Availability, providing meeting summary updates.

Dr. Holmberg.

Agenda Item: Committee Updates - Meeting Summary of DHHS Advisory Committee on Blood Safety and Availability - Jerry Holmberg, PhD, Executive Secretary

DR. HOLMBERG: Thank you.

First of all, I want to say thank you to Linda Smallwood for the years of service too. That's quite an accomplishment. I don't know if I will be able to do 17 years as an executive secretary for the Advisory Committee on Blood Safety and Availability, but in my previous life I was on the BPAC and Linda was the executive secretary during that period of time. So, I appreciate all the hard work there, and enjoy the retirement. She looks so relaxed, and in good health and good spirits.

Also, I'm not Irish, but would like to wish everyone that is Irish or has some Irish heritage in their bloodline, happy St. Patrick's Day. And my mentor -- I saw his wife just earlier today, so I'll to make sure I send an e-mail to him and wish him a happy St. Patrick's Day.

In the next few minutes I would like to update the committee on the activities of the Secretary's Advisory Committee on Blood Safety and Availability. Since I have not been here since last summer to highlight our activities, I thought that this would be a good time to bring you up-to-date with the recommendations that were made from the advisory committee to the secretary, and the action that the department and the agencies have taken.

I think that this will be very beneficial to you as members of the BPAC to hear some of the issues that have been presented to the advisory committee, because many of these issues you have also addressed.

If you have not been to our Web site, I would encourage you to do so. The Web site is posted there.

In August 2004, we took the opportunity to go through various issues that have been previously presented to the advisory committee, and to do an update on progress that has been made. The recommendation that came out of the August meeting was in regards to TRALI. That the committee review the Transfusion Related Acute Lung Injury data, and did not find scientific evidence to recommend an intervention at that this.

But the committee went on to recommend to the secretary to support the expeditious development of standardized definitions of the disease entity, and also the implementation of clinical education and effective surveillance. And also, the modeling the impact of deferral or screening interventions, and also to support to research into the etiology, diagnostic testing, epidemiology, treatment, and prevention.

The secretary's response to that is to continue to monitor the progress of the scientific community. And also, I'm pleased to say that NHLBI has taken on TRALI as a top priority of all non-infectious transfusion complications, and that two institutions are currently supporting investigation.

The next issue was the access to treatment for individuals with rare blood disorders. And the committee noted the importance of this, and encouraged the secretary to facilitate obtaining additional licensed indications for already licensed products, which is a real issue as far as the reimbursement; the approval of products and their indications in the United States for European-licensed products; and also to facilitate developing new products. Also, recommended to the secretary to support research, and also the reimbursement of the use of these plasma products.

The response from the secretary is that the recommendations are being considered. One of the most monumental initiatives that has come out in quite a while is the medical innovation document. This can be located on the Web site that I have given you there. And this is an initiative by the Department of Health and Human Services with all the agencies to try to forward and to move medical innovations ahead.

The FDA has their critical pathways, and also am pleased to say that we are working on various workshops to be able to bring these innovative or new treatments to light, and take action on what needs to be done, next steps that need to be done.

Of course, the bacterial contamination or detection in platelet concentrates and seven day platelets was another issue that the committee has reviewed in the past. The recommendations basically were that the department support through funding, development of bacterial screening suitable for release testing.

And then also that the department consider alternate strategies that could expedite licensure of seven day platelets in less than two years. And also that the department develop strategies to expedite licensure of pre-storage pool whole blood derived platelet components.

The response from the secretary is that these recommendations are being carefully considered, and that FDA innovative regulatory pathway allowing collection of post-approval information on QC data, and also their modified field study has moved the progress ahead very rapidly.

Also, the various agencies within the Department of Health and Human Services have worked very closely with the AABB Task Force, and providing guidance and information to try to move ahead in this area. Several things that the AABB Task Force has done is that over the last couple of months they produced two guidance documents to the blood community, and also conducted a secondary survey of the current status of platelet use in the United States.

And I'm pleased to say too that I hear of a lot of activity with manufacturers, and that I think that the recommendation of trying to move forward in the two year timeframe, we're well on our way, so I think we are making some great progress. And again, I think this is a good example of government working with the blood community to take care of a problem and to move forward with some solutions.

In August we were also asked to look at the hepatitis-B virus minipool nucleic acid test for blood donor testing, and to look at the public health impact of implementing that. The recommendation from the advisory committee was that the committee believes that the comparative expenditure of health care dollars to the general public health would be better served by expanding the HBV immunization program, and encourage the development of a multiplex direct pathogen testing on individual donors.

The response, the secretary felt that the discussion was extremely helpful, and that a status report would be forthcoming. The recommendations from the BPAC, from this committee, and also from the Advisory Committee on Blood Safety and Availability were presented to the internal committee of HHS, the Blood Safety Committee. And when the Blood Safety Committee was presented with the recommendations both advisory committees, they concurred with the FDA policy on hepatitis-B virus minipool NAT test.

The concurrence from the Blood Safety Committee's review of the previous discussion was that the Blood Safety Committee concurred with the FDA policy that: "Current data do not support a recommendation for routine use of the Roche Molecular System HBV minipool nucleic acid test to screen blood and plasma donors. Existing donor screening tests still appear to be adequate, and the new test appears to provide very limited public health benefits at this time. However, public health officials will reconsider possible recommendations for routine donor screening for HBV by nucleic acid tests based on experience with voluntary use of the test, further technology developments, and any other factors that might affect the public health benefit expected from such testing."

In our January meeting we reviewed three different topics. The first topic was the current status of bacterial detection of platelet concentrates, availability and progress towards seven day platelets. Identification of reimbursement issues associated with plasma, and recombinant analog therapy. And the third topic, which was the second day of our meeting was current and emergent infectious pathogens; sharpening our approach for the 21st century to reduce the risk of transfusion transmitted diseases.

Regarding the bacterial blood safety initiative, the recommendation was that the secretary request cooperation of appropriate agencies with blood organizations and transfusion facilities to establish an ongoing program to monitor residual bacterial contamination risk, to provide resources for surveillance of transfusion-associated sepsis, and make additional recommendations as needed to maintain recipient safety.

We, within the department, are still reviewing these recommendations, and I cannot give you at this time, the response from the secretary.

The second topic that we talked about was the reimbursement of plasma derived products and their recombinant analogues. The committee endorsed several principles for the department. And these principles were: plasma derived products and their recombinant analogues should be reimbursed at rates consistent with the true cost, including costs of distribution and administration.

Reimbursement should be sufficient to insure an adequate supply of these therapies; and individual products within product classes should be recognized as therapeutically unique; reimbursement should be provided in different care settings; and lifelong cost of treatment to the individual patient should be addressed in an any pricing structure, including the extraordinary impact of co-payment.

And then committee urged the secretary to support and propose policies or legislation to address the extraordinary financial burden for these patients.

Once again, comments that I made previously about the plasma issues from the August meeting we're continuing to work on, and again, the secretary has not responded to these recommendations at this time.

I also have to point out that we did have a change in secretaries. Sec. Thompson left the department at the beginning of February, and Sec. Levitt was at the beginning of February -- the nomination was approved by Congress.

The next topic that we talked about was on the second day, and this topic was a discussion on emergent infectious pathogens, sharpening our approach for the 21st century to reduce the risk of transfusion transmitted diseases. The overall structure was that we looked first at the IOM report on microbial threats to health. We also looked at an overview of the current blood-borne threats.

We looked at how the system approaches these threats, not only the government system, but also the public sector. And then included in the public sector, we addressed the consumer issues. And then we looked at various case studies. We used West Nile as a model response; Chagas as an unmet challenge; the HIV as evolving challenges, intervention, and donor management; and the HHV-8 as unresolved scientific evidence; and then with the vCJD, how do we handle risk communication.

In the interest of time I don't want to spend too much going through the IOM report, but I would encourage the advisory committee to take a look at the IOM report. It is on the Web site. You can go to the IOM Web page and take a look at that. But there was a very important outcome from that report with all the findings and recommendations there.

Based on our discussion at the Advisory Committee for Blood Safety and Availability, the committee did not make any recommendations to the secretary at that time. The Advisory Committee for Blood Safety and Availability will continue to address the emerging infectious pathogen issue, and we will be discussing this at the next meeting.

Some of the areas that have surfaced that need additional discussion are: surveillance, both known and unknown surveillance; appropriate research; product development; global information sharing; transparency in policy process; and also risk communication.

Thank you.

DR. ALLEN: Any questions? Thank you very much, Dr. Holmberg. Questions from the committee?

Thank you. We look forward to further updates in the future.

The next of our committee updates is by Dr. David Asher, a summary of the Transmissible Spongiform Encephalopathies Advisory Committee meeting.

Agenda Item: Committee Updates - Summary of the TSEAC Meeting - David Asher, MD, OBRR, FDA

DR. ASHER: Thank you. Good morning.

The 17th meeting of the FDA TSE Advisory Committee took place on February 8. Three informational issues and three decisional issues -- the bullets for those are bolded here -- were reviewed, two of those for discussion, and the last one for vote.

First, Lisa Ferguson of USDA reviewed the worldwide BSE situation.

Kate Soldan and Anna Molesworth of the UK Health Protection Agency, CJD Incidents Panel reviewed the UK September 2004 notification of certain recipients of plasma that they might be at increased risk for exposure to the variant CJD agent, and that they should not donate blood or tissues, and should tell their families, dentists, and surgeons to take special precautions to avoid secondary transmissions from their tissues.

Lynn Sehulster of the CDC reviewed what such precautions might be, although CDC currently recommends such precautions only for known CJD patients, and not for instruments exposed to tissues of persons at increased risk.

You may recall from a statement that I read at the last BPAC meeting that factor XI from the United Kingdom was used for a limited time in the US under IND. And that situation was reviewed by FDA's Mark Weinstein, Dorothy Scott.

The promised risk assessment model, which I will summarize briefly in a minute was presented by Steve Anderson, who also reviewed as a separate topic, a draft risk assessment model that we hope will be suitable for addressing the risk from US factors VIII, IX, and other plasma derivatives.

FDA's Pedro Piccardo and Steve Anderson then presented information on variant CJD in France from Jean-Philippe Brandel, who could not attend the meeting.

And finally, Alan Williams reviewed the issue of possible additional deferrals of blood and plasma donors for history of transfusion in Europe outside the United Kingdom on which the committee then voted.

The news on bovine spongiform encephalopathy is guardedly good. The diagnosis of the disease peaked in the United Kingdom at the end of 1992, beginning of 1993. Cases in the United Kingdom continued to drop. They diagnosed 612 cases there in 2002, and only 338 last year, although that number still exceeded the number reported for any other single country.

The 22 other countries' diagnosed cases in the European Union dropped about 40 percent, although 6 countries, which I show here in red, recognized increases in 2004 relative to 2003. It's not clear, at least to me, whether those increases represent better ascertainment or a trend in disease, but they certainly bear watching. We are all aware of the Canadian BSE cases. They were not discussed.

USDA has now completed more than 270,000 tests of brain from US cattle identified as being at increased risk for BSE. They have had 3 inconclusive tests by ELISA. None was confirmed by immunohistochemistry.

There is continued concern about countries of unknown BSE status. The UK customs and excise information showed that meat and bone meal was exported during high BSE years to a number of countries, several of which, like the US, have I understand no record of receiving the material, but it was exported with those countries designated as the country that was going to receive it.

A number of countries in Southeast Asia, not just Japan, and the former Soviet Union are on record as having imported substantial amounts, although that can't be independently confirmed. That also bears watching.

Now, let me turn to variant CJD, which has been, as you know, a major concern of the Food and Drug Administration for several years, because of the potential for regulated products of human origin to transmit the disease. The disease, as we discussed previously, differs significantly in its clinical presentation, histopathology, and epidemiology from other forms of Creutzfeldt-Jakob disease. And our concern of course was heightened by two credible reports of blood-borne variant CJD in December 2003 and December of last year from the United Kingdom. We have discussed those here before.

Concern about the risk of transfusion transmitted variant Creutzfeldt-Jakob disease, even before those two cases prompted FDA to issue revised guidance in 1999 and 2002 to reduce the geographic risk of exposure of donors to the BSE agent. I included in the set a slide summarizing the latest revised guidance 2002. Most of you are familiar with that. I won't discuss it today. Dorothy Scott, who is the keeper of the guidance, is here and any specific questions I suppose we could answer.

Worldwide cases of variant CJD continue to increase slowly to a current of 170, of which all but 16 are or were in UK residents, and only 8 of them had never been in the United Kingdom. Of special interest is a recent report from Japan of variant CJD in a man who had visited the United Kingdom for only 24 days in 1989, which is probably as close to a point exposure as we can expect to see, and implies an incubation period by the oral route of about 12 years.

New cases in the United Kingdom peaked in recognition of new cases in 1999, and deaths in 2000. Those were all in persons homozygous for the gene encoding methionine at codon 129 of prion protein encoding gene, which is a genotype known to increase susceptibility to other forms of sporadic CJD and iatrogenic CJD.

However, troubling new information suggests that as for the other forms of CJD, persons with the methionine valine heterozygous genotype are probably not absolutely resistant to the infection. They may have a lower -- we hope they have a lower attack rate, and they may have longer silent incubation periods.

Nine cases from France -- I think it's really too early to discern any trend in those cases. One of them is still living, two of them were active blood donors.

In case there are any specific questions, we have been over these before, I included in this set a summary slide describing the two cases presumptive transfusion transmitted variant Creutzfeldt-Jakob disease. Both recipients received non-leuko reduced red blood cell concentrates.

The second case was of special interest, because that person who died of unrelated causes during what is probably the asymptomatic incubation period of variant CJD died without CNS pathology, but had accumulation of disease-related prion proteins in spleen and a cervical lymph node, and it's that individual who is heterozygous for methionine valine at codon 129 of the prion protein encoding gene.

Now, let's turn to the Factor XI issue. A total of about 50 individuals in the US were treated with Factor XI concentrate under IND from 1989 through 1997. None of the Factor XI used here was from a so-called implicated lot. That is, no donor actually diagnosed with variant CJD contributed to the source material from which the Factor XI was prepared.

However, because variant CJD has at least a possible prevalence in the UK sufficiently high, it seems very likely that some asymptomatic infected donors may well have contributed. Steven Anderson, with Huong Yang(?) and others in the FDA developed a risk model and preliminary exposure estimates for recipients of the Factor VIII in those studies.

As promised in the last BPAC, Steve presented the model formally at the TSE Advisory Committee, and also as promised, FDA has engaged in discussions with the CDC about approaches to notification and the strategies and the messages that might be offered.

The risks for transmission of infectious agents by regulated products are often estimated and later managed by looking at the risk inherent in the source of the raw materials, the ability of the manufacturing processes to inactivate or remove contaminating agent, and the amount of agent to which the recipient of the product is likely to be exposed per dose per course of therapy per year per lifetime.

And Steve used that traditional model to develop a quantitative or so-called probabilistic risk model, looking at the source of the material, which was recovered plasma, the probability that a given donor to the pool was incubating variant CJD, the number of donations in the pool, and the possible infectivity levels in the plasma. Then he estimated the potential reductions in infectivity that might be assumed for the manufacturing process.

And then finally, at the total amounts of infectivity to which recipients of the product in three typical scenarios, because we didn't have the actual records of what -- we had reason to think that the recommended doses might not have been those doses that were actually used in the IND. So, Steve looked at three typical scenarios that seemed to be reasonable to have expected that the recipients might have received.

Modern probabilistic risk assessments account for biological variability, and especially for uncertainty of the assumptions by assigning to most of the assumptions, a distribution or range of values. Steve used so-called triangular distributions of most likely minimum and maximum values.

Of course the actual values assigned are always open to challenge, but the basic model allows for easy recalculations incorporating different values if people object to the magnitudes of the assumptions that are actually used. It's the computer model itself that is the most important thing.

So, for example based on a recent survey of abnormal prion protein detected in appendices, a survey conducted in the United Kingdom -- more about that in a minute -- Steve assigned a most likely value of two for donors incubating variant CJD who might have contributed to a pool of 20,000 donations, with a range anywhere from 0 to an assumed maximum of 14 donors, which was the greatest number that might reasonably have been expected of undiagnosed, unsuspected donors to have contributed.

Based on animal studies, a broad range of 0.1 to 1,000 animal intracerebral 50 percent infectious doses per milliliter of infected blood with 10 most likely was assumed. But point estimates were used for the percentage of infectivity expected to occur in plasma, 58 percent, and 5-10 fold reduction in infectivity was assumed to account for the relative inefficiency of the intravenous route compared to the intracerebral route, since the preponderance of evidence supports those assumptions. So, triangular ranges weren't used there.

For reduction of infectivity by manufacturing process for Factor XI the assumptions were based on published studies for other plasma derivatives. Steve assumed that anywhere from none to 10,000 fold reduction with 100 fold reduction considered the most likely. And from those, the model generated a series of varying each assumption over the range that has been selected, weighting towards the most likely. The computer model generates a number of outcomes, and those outcomes are analyzed.

The uncertainties of the results must not be dismissed. These are the results of Steve's analysis. You have them in your handout. For example, for scenario 3 in which a subject received 15,000 units of Factor XI, there might have been a mean estimated exposure to 0.28, 50 percent infectious doses, which suggests the chance that 14 percent or 50 percent of 0.28 individuals that received the dose, that dose of Factor XI from the United Kingdom might have been exposed, and some of them might have been infected with variant CJD.

However, the animal data may not accurately reflect the potential for variant CJD infection and illness in humans. Even animal vary in their susceptibility from one species, one strain to another. And these uncertainties are no less important than the computer generated most likely risk estimates.

CDC pointed out at the meeting that the BSE outbreak was at first identified in 1986, and peaked early in 1993, about seven years later, which should have been ample time for human plasma derivatives to have been implicated in transmissions of variant CJD. Yet, there has not been a single case in the United Kingdom attributed to exposure to a derivative, which gives some reason to hope that these estimates may be too high.

Quantitative probabilistic risk assessments have several advantages. First is transparency. If you distrust the assumptions, you are free to suggest other assumptions, and those can easily be incorporated. In response to suggestions at the meeting, Steve is now adjusting the prevalence assumptions to allow for differences in subject ages and years studied, and for the years in which the subjected donated.

Data gaps are easily identified, and by holding magnitudes of the assumptions constant, except for one, and then systematically increasing the magnitude of each assumption separately, and examining the estimated risk outcome called sensitive or importance analysis, the critical drivers, the most important elements of risk can be identified.

For data gaps, models of variant CJD in the United Kingdom in the past eight years have predicted a widely different number of cases, anywhere from a minimum of 29 -- there have 154 already, so that can't be right -- to 13.7 million, which also seems improbable, most recently from a few hundred.

But there is now actual empirical survey data available suggesting that somewhat more than 100 person per 1 million, 1 per 4,000 living in the UK may be in the preclinical incubation period of variant CJD, and several people there may eventually get the disease, although several authorities at the meeting suspect that those figures may be too high. It seems at least prudent to include them in any risk model.

Some data gaps are slowly filling. We have mentioned the incubation period 9-12 years seems reasonable for food-borne; 6 years is clearly an incubation period in one of the blood-borne cases, but of course there is more that we don't know. We don't know the infectivity levels in the blood during the incubation period. We know that blood is infectious for at three year before onset of symptoms, but we don't know how long before that. And all that would be important in estimating risk more precisely.

Two major drivers or risk, one the overall prevalence of infection, and second, the reduction by infectivity. I just want to mention briefly in this regard that other forms of CJD appear to be quite different. Many years of experience with sporadic and familial CJD in this country have failed to detect any transfusion transmitted cases. So, convention CJD seems to be quite different in its biology.

As you know, the FDA has recommended safety policies to reduce the risk that a donor might be incubating CJD of any kind, while not deferring so many donors as to compromise the blood supply. And we of course acknowledge that the policies cannot possibly eliminate all conceivable risk.

In October we offered to the TSE advisory committee, a question of considering whether the precautions were adequate to protect against variant CJD, whether they were still justified, and whether enhanced precautions are needed. And I list here the possible approaches to reducing the risk.

One, eliminating donors how had potential dietary or pharmaceutical exposures to BSE agent, and the other deferring donors who had exposure to human blood or blood products from donors who themselves might be incubating CJD. Those are the approaches that can be taken to enhance the risk reduction.

Alan Williams both in October and at the February meeting reviewed the estimated reductions in risk with the loss of donors that might result from three new policies which I summarize here. As usual, reducing the time that a suitable donor spent in the UK would reduce the risk the most, but would cost many, many donors.

I would add if we attempted to defer any donor who spent any time in the UK after 1980-1996 -- and Japan just implemented such a policy, we estimate that we would lose 23 percent of current donors. And if we tried it for the rest of Europe, it would be total loss of 36 percent. So, it's beyond the range of the possible.

Deferring donors transfused in France would reduce the risk slightly at a cost of 1.4 donors; the rest of Europe a cost of 3 donors per every 10,000. There is a problem in that European blood from US licensed establishments in the Netherlands, Germany, and Switzerland through the New York City Euro Blood Program or the major source of exposure to European blood in the US population, those recipients cannot be traced, and they could not be deferred without practically deferring all the other recipient transfused donors in the New York City area, which is a doubtful prospect. Alan Williams discussed it at the meeting. I won't discuss it further today.

So, in October we asked the committee, without offering specific options, to consider whether the deferral policies currently in place are still justified, and whether enhanced policies are indicated. They unanimously endorse the current policy, but did not recommend any enhancements in the October meeting.

However, in associated discussions, members seemed concerned about the growing number of cases of variant CJD in France, and the higher BSE exposure risk there because of major imports of UK beef products. Five to 10 percent of the French beef during the late 1980s to early 1990s, 60 percent of the total UK beef exports went to France. They had a high level of exposure. They have had the most variant CJD cases outside the UK.

So, the committee expressed some concern about the growing number of cases in France, and the higher BSE exposure risks there from the imports. And at the February meeting we offered them specific options to vote on. They voted overwhelmingly to recommend deferral of donors transfused in France, 12 to 3, with 1 abstention.

And they split narrowly on whether to defer plasma donors with that history, 5 to 4 with 7 against it, but with an unusually high number of 4 abstentions. No member felt that the deferral of blood or plasma donors transfused elsewhere in Europe outside the UK and France was currently advisable.

Let me close by saying that we realize that these deferral policies are inherently wasteful. We all look forward to the day when BSE and variant CJD are completely eradicated. Before that happens, it would be very useful to have effective screening tests to identify those donors, and only those donors who are infected, and devices that will remove infectivity from blood components.

Unfortunately, at the moment there is no FDA licensed technology to do either, but we look forward to hearing more about some of the exciting new technologies under study, and we will look forward to receiving applications for licensure of those methods as soon as possible.

I'm sorry that I ran over.

DR. ALLEN: It was indeed a lengthy and very involved discussion. I think the votes that Dr. Asher presented reflect the uncertainty of exactly what direction to go in each instance.

Comments or questions for Dr. Asher on this issue? Okay, thank you very much.

We'll move on to the next presentation, Dr. Alan Williams' update on West Nile virus guidance.

Agenda Item: Committee Updates - Update on West Nile Virus Guidance - Alan Williams, PhD, OBRR, FDA

DR. WILLIAMS: Thank you and good morning.

At this committee's last meeting in October you discussed extensively available data related to West Nile virus natural history and the testing programs, and made recommendations with respect to policy issues. What we wanted to do today was update the committee on FDA's current considerations, which should be interpreted as considerations which are serving to help frame policy development.

The thoughts here should not be really interpreted as FDA recommendations at this point. Those would only come with publication of final guidance. And we anticipate that with West Nile virus, that guidance, when issued, would issue as draft with the opportunity for a formal comment period.

Previous FDA recommendations were published in October 2002, and then revised May 1, 2003. Key recommendations included deferral for West Nile virus infection of 28 days from symptom onset, 14 days post-symptom resolution. And this was based on historic data with respect to the longest known incubation period in immunocompromised individuals.

In addition, with the revised guidance FDA recommended a deferral based on the donor self-report as a result of directed questions for report of a fever with headache in the past week, and this would also constitute a deferral for 28 days. And this was instituted based on limited case control data from some of the earlier known transmissions from donor to recipient of the infection.

Observational data, most of which was presented and reviewed at the last meeting, related to donor screening by West Nile virus NAT testing and follow-up studies conducted under IND. To date, the maximum observed West Nile virus viremic period in blood donors is 49 days. The observed or predicted range based on 1 model is 6.5 to 56.4 days, and this is projected to include 99 percent of the West Nile virus infected donor population.

It is known that low level viremia in the late stages of infection generally is co-existent with IgM antibody to West Nile virus, and the infectivity of this combination is currently not known definitively, although I think an important observation is that no post-transfusion West Nile virus cases observed to date have been from IgM positive donors.

The NAT testing for the major portion of the year is conducted by minipool testing from 6 to I believe 16 samples. But there is the option to conduct individual donation testing. This improves the sensitivity of the NAT donor screening, but currently is not feasible year round. And as discussed again at the last meeting, it is conducted on a targeted basis at the peak epidemic periods for different geographic areas.

Donor screening for West Nile virus-related symptoms introduced prior to widespread West Nile virus NAT screening is of questionable predictive value. The appropriate studies haven't been done to really assess formally the predictive value of the question. But data related to the relationship of the timing in the epidemic, and the deferrals related to this question do not seem to imply a close relationship.

The geographic focus, timing, and extent of what West Nile virus will do in 2005 is not known, but it is anticipated that there will be continued epizootic activity, including in those areas that were previously affected.

At the committee's last meeting, October 22, 2004, there were recommendations made by vote. One was based on the observation of the potential for a longer viremic period, to extend the deferral period from 28 to 56 days. And not a vote, but comment and discussion related to encouraging data collection to determine if 56 days is in fact sufficient to rule out viremia in rare instances.

There was discussion that at the 56 day period it may appropriate to conduct West Nile virus individual donation testing prior to re-entry, but that also a reasonable period of time beyond that, it might be appropriate to consider re-entry without specific testing at the time of re-entry, but simply the normal donor screening test that is done.

The committee also, by an ad hoc vote, recommended discontinuation of the previously recommended question regarding donor reports of fever with headache in the past week, and encouraged that better case control data may determine if there are any symptom clusters in the early prodrome or acute phase that may be more predictive of early West Nile infection. FDA currently considers that this recommendation of the committee is appropriate.

Additional considerations include 90 day deferral. And this is simply looking at it the other way. A 90 day deferral with the potential for re-entry at 56 days. Those donors with diagnosed or suspected acute West Nile virus illness or infection. This is based on 90 days after onset of illness or diagnosis, whichever is the latest date. Also, those donors with suspected post-donation West Nile virus illness, 90 days post-onset of illness. And donors who may have transmitted West Nile virus infection, 90 day deferral from the implicated donation.

And as mentioned, there is consideration that re-entry may be possible after 56 days based on negative individual donation testing NAT on a sample or a sample from a donation at the 56 day period or beyond. And that re-entry would be feasible without additional testing after a 90 day period.

In terms of product management, in cases of diagnosed West Nile infection or illness in a donor, consideration is for retrieval and quarantine of in-date products 14 days prior to onset of illness, reflecting the 2-14 day likely incubation period.

And 90 days after diagnosis or onset of illness, whichever is later, reflecting again the potential viremic period plus an additional margin of safety, with an exception for products collected more than 56 days after diagnosis or onset of illness in the donor, when that donor has been tested and found negative by individual donation testing NAT.

Product management for donors who are potentially associated with transfusion transmitted West Nile infection. This is defined as donors of suspect donations received by recipient up to 56 days prior to recipient West Nile infection. And the consideration is for retrieval and quarantine of other donations by a potentially associated donor 56 days before, and 56 days after the suspect donation, essentially bracketing the viremic period on either side of the potentially suspect donation.

Production management in cases of undiagnosed post-donation illness in individuals who may have been exposed, as in the previous recommendations this would appropriately be based on medical director judgment regarding the product quarantine and retrieval in light of the timing of the illness, its relationship to West Nile epizootic and epidemic activity in the geographic area, which would be typically May 1 to November 30, but could be one month in either direction depending on the status of the epizootic.

When conducted, quarantine and retrieval of in-date products would be 14 days prior to, and 90 days after onset of donor symptoms, with again, an exception for products collected more than 56 days after diagnosis or onset when a donor has been tested and found negative by IDT NAT. Product retrieval and quarantine are not recommended for source plasma, recovered plasma, or source leukocytes that have been pooled for fractionation.

And then finally, with respect to notification of prior transfusion recipients about the possibility of West Nile virus exposure, those units considered relevant are those that are collected 14 days prior, through 56 days after the onset of diagnosed West Nile virus illness in a donor. And this is definitive West Nile virus diagnosis, not non-specific symptoms or history of extensive mosquito bites for instance.

Or another situation, units collected from a donor 56 days prior to 56 days after a donation from that donor is identified uniquely as the likely source of a West Nile virus transfusion transmission. Consideration is that establishments should consider tracing records, and notifying transfusion services of relevant units.

Again, this represents current considerations of the agency, and recommendations that will be issued in draft with a formal comment period.

Thank you.

DR. ALLEN: Thank you, Alan.

DR. KATZ: Three things. First, is current considerations the same as current thinking? A small point, and a little bit facetious.

Alan, as the temporary non-voting and other qualifications that I might not think of industry representative, I have been asked to express a little frustration. And so, since I'm not my own agent at this table, I will do so.

We have no final guidance, no final draft at this point. It's been five months since the discussion at BPAC, almost the same amount of time it took us with the test manufacturers to develop, install, validate, and implement nucleic acid testing. And people are very, very frustrated about that.

We can implement a new nucleic acid assay almost as fast as the agency can provide us what we think should be fairly straightforward guidance. And our problem with that is that it's too late now to apply whatever guidance occurs in the next several weeks, or whatever the interval is for this season under the CGMP processes that we need in order to change the direction of the Titanic, for all intents and purposes.

And also, that we have a continued loss of donors for what we recognized at the time were unvalidated donor history questions, and which we know now from the Red Cross and blood systems data to be ineffective donor screening questions.

So, I think I have made the point that I want to make, and I don't know that I expect a comment. I think we all understand the legal constraints under which FDA operates in terms of talking about the guidance ahead of time. The real issue here is why does it take so long?

DR. ALLEN: That was two points. Was there a third?

DR. KATZ: There was, but I have forgotten what it was.

DR. WILLIAMS: Okay, I'll comment briefly on the first point, maybe very briefly on the second, and Kathleen Swisher from our policy office is here. She may either want to correct me or add something to that. Current thinking is a formal term that is part of GGP. And current thinking, to my understanding, is the thinking that is reflected in a guidance document. Current considerations reflect the considerations that go into policy development before that stage is reached.

In terms of the timeframe, I think the only comment that I can really make is there is a lot of process involved in developing policy, and it competes with many other things that are underway within the office and the center. We are very sensitive to the timing of the upcoming epidemic. The peak period of epidemic still is anticipated to be in the August-September timeframe.

I guess I would be interested to hear why next year would necessary be necessary for a full implementation of policy. But we are sensitive to the concern, and there just is process that needs to be followed.

DR. KLEIN: I would just like to have a follow-up to that. Since we have some relatively new data, at least on the incubation period, and when I say relatively new, as of last October, and on the relevance of some of these questions. Is it not possible for the organizations to put together a policy that the FDA would concur with which would make more sense than some of the current policies that we're using?

DR. EPSTEIN: Well, the agency is always open to suggestion. We call them unsolicited proposals. However, the process for us to establish them as FDA policy is the same under good guidance practice. Namely, we would have to publish it for public comment, simultaneously by all interested parties.

So, it doesn't buy a lot of time at the FDA end. On the other hand, if in this interval the industry went forward with voluntary policies that appear to be in no conflict with what the FDA is considering, probably things will work out fine.

I would also remind you that a year ago we published guidance on implementing the question. I guess that was for 2003, the question about fever with headache. And our guidance published in May for implementation by June, and I know it's a painful exercise, but the industry was able to implement something de novo. Whereas, what we are saying here is basically to back off. So, I would think that by May or June it would be feasible, although I understand it's painful.

DR. KATZ: It's extraordinarily painful. And I'm not convinced even at my center, just a little tiny center out in the middle of nowhere, that those emergency questions where implemented under what I would call CGMP criteria. And so, we would prefer, and particularly if it's not an emergency, to do it very carefully with all the appropriate training and validation.

Would this be an appropriate -- for example, the donor history questions be an appropriate application for a variance, for example, at this early date?

DR. EPSTEIN: I'm not sure how that changes things, Lou. You are saying can you ask now under a variance, to drop the recommended donor question? The answer is yes. But I think that FDA is very mindful that you need to see as soon as possible, at least a draft guidance. And so, we are doing everything we can at the present time to expedite the GGP process. So, I think you will have it in your hands very soon.

And although a draft is a draft and it's not final, mind you final guidance is also not binding on the agency or the industry. So, I think there will be a lot of benefit to see what the agency is currently thinking in the form of draft guidance, and that should happen quite soon.

DR. ALLEN: Other comments or questions? Yes, would you identify yourself, please?

MS. STRAMER: Susan Stramer, American Red Cross.

I would like to complete some of Lou's thoughts for points maybe that he didn't. And first, let me address Dr. Epstein's point. For us to make simple changes, these changes have to go through very well controlled and defined processes. And our documents, including our consent forms through our donor notification letter, and all those processes only come up for renewal at certain time sequences.

So, we have our priorities set, just like the FDA or any other organization. And when there are major changes which have just been introduced in the last couple of days, it's impossible to turn the organization on its ends to have those changes implemented. So, even by the end of the season we have already done investigations, and they can't occur. The next time some of these documents are up for revision won't be until the September timeframe.

So, something as simple as a slam dunk as you would think, eliminating the headache with fever question, the earliest that could be done is in a revision that will only come out in September. For consent forms to be changed, which link the way we do donor follow-up now to reinstatement, not only would those have to be changed, go through the IRB.

But once recognized as formal documents and changed through our system, would have to be trained in tens of thousands of individuals and collection sites throughout the United States. So, clearly these types of changes are very difficult to implement in a very short period of time, which we have now.

And then regarding 56 to 90 days, if we collect a sample for individual donation testing post-56 days, that isn't equivalent to re-entry at 56 days. Donors will not come in on 56 days for a subsequent follow-up sample. And by 56 days, they don't care about a follow-up sample. We have lost the donor at that point to follow-up.

So, a program like this would be tantamount to eliminating follow-up in an investigational protocol. So, not only do we have to bring the donor in at some time post-56 days, by the time you collect consent and collect the sample, send the sample for testing, turn around the test results, notify the region, the region notifies the donor, you are at 90 days.

So, why in the world would we have two tiers? We would just automatically defer everyone to 90 days. And data on follow-up would terminate.

Thank you.

DR. EPSTEIN: Our current consideration is that the donor would be able to give a unit on or after 56 days, provided that that donor has an IDT test. So, we are not thinking of an offline requalification test. We are thinking that the donor could come back as per usual after 8 weeks of donating whole blood. And that as long as the donation is qualified by IDT NAT, that that is a suitable collection.

So, I think that the cycle of testing interpretation, re-entry, and requalification is not really what FDA has in mind. Now, I understand your point that it may be operationally simpler to defer for 90 days, but we are not precluding a collection at the normal 8 weeks. We're just saying do IDT NAT on that unit.

MS. STRAMER: And therein is the rub. For us, when we collect in our system, 22,000-25,000 tubes a day, it's impossible for sample management to flag that sample and say, this one sample goes to individual donation testing while the rest of the units go for pooling. That's operationally problematic, would led to error, and these are all manual processes. In our system, manual processes don't get approved, because they don't work. So, we would be back then to 90 days.

DR. ALLEN: Any other committee comments or to add to that discussion or questions?

I think this is a very important issue in terms of process. It does have a real impact on the blood collection agencies. On the other hand, we want to assure the safest blood possible, but there are lots of practical considerations. I know that the FDA is working on this. The committee has not been asked a formal question on it, but I'm sure that there probably is agreement that we would encourage the FDA to continue to work on this expeditiously with the blood collection industry to try to resolve the issues.

DR. LEW: Just a question. For the 56 day NAT testing, wouldn't that be data that would be quite helpful to us? Because a lot of this data is extrapolated from limited data already. And I can see in a process it might be easier in a large blood collecting facility to go to 90 days. But if you think of it in terms of more a research project, which may provide us good data to help decide exactly when a patient can come back, for the whole group it might be more useful.

Again, I am concerned all this is extrapolated. And we really don't know if 56 is appropriate or not. It's the best data we have, and it's reasonable to use it.

DR. ALLEN: Those are very good points. The difficulty is that nobody is paying for this as a research study. Nobody is paying to have the donors come back right at 56 days plus or minus 48 hours, whatever it is. The donors come back when the donors come back. If you have done something that they consider to be off-putting in some way, most donors don't come exactly 56 days after their previous donation. I think it's generally a much longer period of time.

Do you want to address that, Dr. Katz?

DR. KATZ: Well, I think that Dr. Allen made the important points. That while we love to do operational research, we are not being paid under most circumstances to do it. And the complexities of flagging these donors and bringing them back are substantial in operations that are trying to draw hundreds of donors a day in buses, in church basements, in fixed sites spread over eastern Iowa and western Illinois or North Dakota, as Dr. Bush will tell you.

So, while it's a great idea and we would all love to have the data, the business is to get blood for patients, not research, and in most cases it's operationally difficult.

DR. LEW: Well, I guess just in response, what I heard earlier in other scenarios, for let's say variant CJD, we're losing a lot of good donors for probably some low, low risk. I don't know, is it possible we go to this 90 day automatically, again how many donors are we going to lose needlessly if it really needs a shorter time period. If there is enough data and people are content, 90 days is good, fine. But I'm not sure I see that data.

DR. BUSH: We have follow-up data between Red Cross and blood systems on about 600 West Nile viremic donors who have enrolled in follow-up. That's probably three or four times the HIV and HCV positive who ever enrolled to give longitudinal data. So, it's actually an enormous data set.

And the follow-up demonstrates one donor who had viremia detected, one of 5 reps at 49 days in the presence of high level antibody. That's the outlier case, that combining the data and modeling out a 99 percent inclusion bound gives you the 56 days; extremely conservative analysis modeling. Which to me, 56 day deferral is reasonable, but the extension to 90 days, and the algorithm that is suggested here is problematic.

We are planning to do more aggressive serial follow-up, including a variety of additional immunology studies, et cetera, under funding from CDC this year. Those donors, because they are going to be sampled weekly, in fact we'll be drawing more blood from them. We will be drawing the equivalent of a unit of blood during the follow-up studies. So, they are going to have to be deferred for two more months anyway. So, they will be deferred for four months.

But the other donors who don't qualify or don't enroll in this study, our decision is that they are going to be automatically deferred for 90 days. That we just can't put into place the logistics of trying to get a sample at 56 days or a donation. So, we're going to try to get more data in a limited number of donors in a more sort of careful, serial context. But I actually think there is a huge amount of data now relative to anything we have ever had, and the 56 days is a reasonable deferral period.

DR. ALLEN: Thank you. Any further comments or questions?

Okay, we'll move on to our next update topic, the critical path initiative workshop summary. There are three presentations, first by Dr. Kathryn Carbone.

Agenda Item: Critical Pathway Initiative Workshop Summary - A. CBER Overview - Kathryn Carbone, MD, OD, CBER, FDA

DR. CARBONE: Good morning. As the Associate Director for Research for CBER my task today is to sort of give a very brief introduction of what the Critical Path Initiative is, and that will be followed more blood-specific oriented talks later.

The Critical Path Initiative was conceived in the office of the commissioner through work done by Janet Woodcock and others to address the issue of the science of product regulation. In other words, to identify why the issues of medical product applications were dropping in an era where of course basic biomedical research was surging.

And as you can see, despite many years of basic biomedical research and phenomenal advances in technology and novel findings, the ability to translate some of this research into actual medical product use seems to be somewhat in question, at least something we should be addressing.

This is a slide that just shows the applications to CDER and CBER that have been declining over a time of increased spending and really activity and really tremendous findings in basic research.

So, what seems to be missing? Well, it turns out that as a science-based regulatory agency, data-driven if you will, the question is are the availability of the appropriate data with which to make the regulatory decisions of efficacy, safety, manufacturability, and then all the unique areas that concern the products that CBER is involved with including the blood industry?

And as you can see, the basic research, the holding up the vial of vaccine and saying eureka is really just the start. That's a candidate, and that candidate needs to be evaluated, tested, studied, and turned into a valuable and reliable medical product.

And as you can see, the prototype design or discovery may be just the beginning, and that preclinical development, the toxicity testing, develop the tools to do appropriate toxicity testing, characterization of the product, et cetera, not to mention the long clinical development of the product really are a tremendous resource and time activity, and yet much less attention has been paid to the science of this stage. Then the application is entered, and of course the FDA is provided with information to evaluate.

So, the three dimensions of the critical path, and obvious to most people in this room, but not so obvious necessarily to those in the basic research realm as well as the public is that the three major areas of concern at the FDA is that we have the scientific tools and the scientific knowledge to make an appropriate assessment of safety, typically through pre-clinical studies followed by clinical trials.

Proof of efficacy. Does it work? Does this have a medical benefit? And then an area which is very interesting and fairly unique to industry and the FDA's concerns, and that's the matter of industrialization. Something that can be made in a small quantity on the bench may not translate into a large quantity or manufacturing of a consistent product.

And the failures at a large scale manufacturing level really mean a tremendous amount of money has been invested in product development, which may end up as a relatively complete waste, because the product cannot be manufactured in a scale to be a viable medical product. These sorts of things are very important. And the proposal through the FDA's Critical Path Initiative was that not enough attention is paid to these steps in the scientific process.

So, the proposal is to do something which the FDA has been working, CBER has been working on for many years, which is to try and apply new science to improve the predictability of product development, not just discovery. And that to develop and support the development of scientific tools and knowledge, so that we can more efficiently assess and more accurately assess these candidate products.

Now, the question, is the outcome of this -- the hope is that by reducing the cost, and focusing the resources, and a concept that is not always well understood outside the actual product development area is the concept that early failure of a product is a good thing, and late failure of a product is a very bad thing.

So, the question is how can we earlier and more efficiently identify the products that will end up being good products, and prove to be safe, efficacious and manufacturable at an early stage in their development? And to identify the products that will not have those features at an early stage, and thereby increase efficiency?

The other thing is by focusing on the science and the information that we can use to make these determinations, we hopefully will find a more quantitative and less empiric approach to product development.

Therefore, the FDA has made an attempt to bring some attention to this through workshops that actually CBER has held last year, and articles and talks with all sorts of sponsors from patient advocacy groups to industry to academia to try and highlight this issue.

Why is the FDA at all interested? Well, we have, we believe, along with other people who are experts at developing products, and because we are involved in the review of the product, we can see the successes, the failures, and the missed opportunities for entire products as a whole field.

That guidance documents, the translation of the science into as we were discussing, I think it's a pretty clear message that the desire and the importance of guidance documents, the translation of this science into public information that can be used generally across the field for developing better products would be quite helpful. And that we see a role with the FDA of convening and coordinating these sorts of activities.

Now, without stepping outside of course, the proprietary bounds, the FDA members and staff have a special, unique knowledge of areas that seem to be lacking or gaps or obstacles, or if you want to be positive about it, opportunities that solutions that help move entire fields forward.

And I just have here a little diagram of the FDA process. For example, amplification comes in. It's reviewed, and several areas are identified, and responses are sent to the sponsor for additional data, et cetera. But in some cases what is needed is not in anybody's purview. There is a total absence of the ability to test something, the ability to standardize something.

For example, you have many serum factors or plasma factors, clotting factors that are difficult to characterize, difficult to measure. There are issues involved with recombinant products, and all these sorts of tests, assays and understanding may be lacking. And rather than regulate in a vacuum, the ability to identify across entire product categories, areas that are critical, blocking if you will opportunities to expand and make regulation more efficient can be identified as part of the scientific process.

Now, we are not talking about the FDA doing all the research on every issue. What we are talking about is the FDA playing one of three roles. The FDA would like in this sort of initiative, if this was possible or this was allowed, to send out information, and at some point support, if it were to come to this initiative, outside the FDA to identify, to educate, to bring areas that are not so expert in the product development issues, and support that kind of science externally.

Also, the FDA could play a role as in collaborations with intramural FDA scientists with outside scientists to work together to solve some of these problems. And of course to strengthen the internal intramural base of the FDA science so that we can play a role. And you will hear from my following colleagues, some examples of some opportunities and some successes with FDA intramural collaborative science.

Then this science of course, when it reaches the public domain, because this effort is a public activity, we obtain public input. It's discussed, and then this information can feed right back into the regulatory process. So, unlike many diagrams you see, the science is not a separate item to the side of the regulatory process, but is actually an integral part, especially with the complicated and difficult products that CBER deals with.

So, I'll end with some possibilities. Of course what I'm talking about is an unfunded initiative. So, no promises, but some of the areas that we think as a center are very important to address are the development of well characterized cell banks for the use and manufacture of some of biological products. It's quite a different approach when you are talking about manufacturing when you have actual living cells that many of these products have to be made in.

Characterization of cells therapies. That will be stem cell or mature adult stem cells or embryonic stem cell therapies. And the links to standardized outcomes. Those would be the bench to bedside, if you will. Trying to identify what makes an embryonic stem cell or an adult stem cell that is quite D-differentiated, how can we tell where that stem cell is going to go, what it's going to become, and reliably predict the product, characterize the product?

Methods and validations of pathogen and activation for blood; plasma tissues and other products; multi-pathogen and rapid detection methodologies for biologics; improving the ability to store blood and blood products and tissues; and other activities as determined by stakeholders.

Again, we had a meeting in October of the entire CBER product responsibilities with our stakeholders in critical path, and these we have put out in the public domain. We now have some of the statistical discussion published. We'll be publishing the general meeting summary coming up. So, we are putting this information out.

But in addition, we are working with stakeholders to try to hone down in certain products, and develop some public information and interest in approaching product development. For example, in the end of May we will be having a DoD an innate immunity critical path workshop to talk about biodefense products. And we'll be addressing in an open forum, the issue about these novel products, many of which are not yet licensed, or categories are not licensed. And how we can move that forward, develop some consensus. And we'll be continuing these kinds of efforts as we can.

That's it. I'll be happy to take a few questions, and then we'll proceed to more blood-specific issues.

DR. ALLEN: Thank you, Dr. Carbone.

We've got two more presentations in this series. What I would suggest is that you ask questions just for clarification, and then we'll come back and let the committee ask all of the speakers for broad overview questions. Any clarification questions?

All right, thank you, Dr. Carbone. We'll move to the second presentation, OBRR summary by Dr. Paul Mied.

Agenda Item: Critical Path Initiative Workshop Summary - B. OBRR Summary - Paul Mied, PhD, OBRR, FDA

DR. MIED: Thank you, Dr. Allen.

I'm going to talk very briefly this morning about what FDA's Critical Path Initiative means in the blood area. And I'll summarize the blood session that was held during the Critical Path Workshop in October. Then I'll tell you about where we see critical path opportunities in blood and blood products.

Now, Dr. Carbone showed you this. Here is the critical path for medical product development. It extends from prototype design and discovery through pre-clinical development and clinical development, to market application, to FDA, and product approval and launch.

This is the critical path challenge for blood products: to find critical path opportunities to improve blood product safety and efficacy, and minimize disruptions to the blood system.

As Dr. Carbone mentioned, CBER held a public workshop in October 2004. The title of the workshop was, "From Concept to Consumer: CBER Working with Stakeholders on Scientific Opportunities for Facilitating the Development of Vaccines, Blood and Blood Products, and Cellular, Tissue, and Gene Therapies."

There was a panel session that focused exclusively on opportunities for FDA and industry to work together during the critical path of blood product development and commercialization. Now, this panel consisted of representatives from blood and plasma organizations, from academia, and various consumer groups. And in the way of overall communication, the panel identified opportunities for FDA such as to increase communication with industry, and to issue guidance documents faster.

Now, it was felt that there should be an information sharing partnership between FDA and industry. And the creation of new forums, new types of meetings for example for information exchange between FDA and industry. And we are well aware of the industry's need to get guidance documents in a timely manner, and we are actively looking at ways to expedite the internal review and clearance process for guidance documents.

Establish a mandate. It was felt that rather than just to allow market forces and market-driven product development to occur, there is a need for FDA to establish a mandate for public health on a particular issue. However, I should point out not where the technology does not exist. Mandates would confirm the agency's thinking on market need, and would in effect act to bring more technologies to the market.

FDA should develop consistency with other agencies and regulatory guidelines. We should work with agencies nationally such as HRSA and CMS to achieve coordination, harmonization, and partnering for consistency of standards, for example, in patient identification and in bar coding. And internationally with regulatory agencies and groups developing standards.

During the research and development phase the panel said for FDA to set standards in advance of product development. The setting of standards by FDA will give industry defined targets for product development. Analytical standards, physical standards, and process standards as targets will help industry proceed in the right direction. For example, in bacterial contamination in blood products, blood donor testing panels, and blood cell survival times.

It was felt that FDA should help with research and product development in specific areas such as gene chip and nanotechnology for direct detection of infectious agents and blood products, including viruses, bacteria, and prions; prevention of and/or detection of bacterial contamination in blood products; pathogen inactivation; animal models and tests for prion diseases.

Also, FDA should help to reduce product development time. FDA and industry should access European technologies, for example, automation of blood component production to improve blood safety. FDA should facilitate licensure of blood components with data from Europe.

During clinical development, the panel said FDA should set objective, firm, approval requirements. These will help industry and product development prior to submission of an application to the agency, such as FDA did at BPAC in stating requirements for licensure of a test for Shagosis(?).

FDA should provide more guidance on the structure of clinical trials. FDA should make better use of alternatives to clinical trials. For example, use pre-existing data to make rational regulatory decisions. For example, using the epidemiological data to support the possibility of testing for West Nile virus only during the appropriate season, or testing for Shagosis only in different geographical regions.

Allow the use of model systems and surrogate endpoints for safety and efficacy. Use risk management strategy, for example to reduce clinical study requirements for blood components in the setting of comparatively small risk.

And finally, in the area of FDA application review, the panel said FDA should accelerate product reviews. Now, the FDA agrees that any reasonable methods for accelerating the review process should be implemented. And to this end, the agency has committed to meeting goals established in the 1999 Prescription Drug Users Fee Act, PDUFA, and the 2002 Medical Device Users Fee Act, MDUFA. And we are actually meeting or exceeding those goals each year.

One specific thing I can tell you that we are doing to accelerate products reviews is during the review process for a 510(k) or a BLA or a PMA, we have been focusing on conducting interactive reviews in which questions are posed to the sponsor, and problem-solving occurs early on in the review cycle.

It was felt that FDA should facilitate the introduction of new donor screening tests. FDA should continue to use mechanisms such as testing under nationwide INDs to implement donor testing for new markers, for example as was done for NAT testing for HIV, HCV, and West Nile virus.

FDA should allow improved access to FDA panels to speed access to new assays. Also, FDA should accept the use of alternative measures to validate product or process changes. As I already mentioned, allow the use of model system and surrogate endpoints for safety and efficacy. Ask for comparabilities studies or analytical studies to obviate large clinical trials to validate these changes. For example, in the immunogenecity of Factor VIII.

Imaging technologies could be valuable as surrogate endpoints for quality of blood components at or near expiration. FDA should develop statistical approaches in phase IV evaluations. And also in phase IV evaluations it was felt that FDA underutilizes risk management strategy. The agency needs to evaluate risk-benefit ratios in the approval of product or process changes.

So, in summary to all this, CBER is listening to comments and recommendations such as these, and we are thinking creatively about solutions in these areas, some of which will involve our stakeholders' help. FDA will be summarizing and publishing this discussion on critical path for blood and blood products and similar discussions that were held for the vaccines and cellular tissue and gene therapy panel sessions in the scientific journal and the trade journal for comment soon.

Let me give you one recent example of critical path success in the blood area. It's an excellent example of how FDA can work closely with industry and be a facilitator at various points along the critical path to bring products to market to benefit the public health.

To help to make donor screening for West Nile virus RNA a reality, in August 2002, FDA issued an alert to blood establishments to exercise vigilance to exclude potential donors with symptoms. In September 2002, FDA issued a call to industry to rapidly develop blood donor screening tests. And in October 2002, FDA issued a guidance document with recommendations to prevent donors with symptoms from donating, and to manage implicated products.

We sponsored a public workshop in November 2002. We discussed various test issues, and we set standards in the development of West Nile virus donor screening tests at various public meetings including BPAC. And we maintained a dialogue with kit manufacturers during the development of their tests, and that dialogue continues today.

We approved large scale studies to be carried out under different INDs at many sites nationwide. And we have been monitoring the epidemiological data and the outcome of nationwide screening of the blood supply in minipool format that was initiated in July 2003, and the triggers for switching to individual donation testing depending upon the rate of positivity in different areas of the country.

At FDA, we are at work developing reference materials and standards that companies can use to validate their tests, and animal inoculation studies ongoing to evaluate the infectivity of West Nile virus at low tider in blood.

Lastly, here are some critical path examples of opportunities for research and development for blood and blood products that we see for the future: the evaluation of diagnostic and donor screening tests for TSEs; development and validation of multiplex NAT tests to detect bacteria and parasites in blood; nanotechnology, a novel technology for blood screening and diagnostic use also; development of a valid animal model to predict immunogenecity of Factor VIII products; and development of standards for plasma-derived products such as alpha-1 protease inhibitor.

These may be fruitful areas in which FDA can focus its efforts to work with industry at different points along the critical path.

Thank you.

DR. ALLEN: Thank you very much, Dr. Mied.

Any questions for clarification to Dr. Mied?

Okay, we'll move on to the third presentation, the clinical trial design by Dr. Mary Foulkes.

Agenda Item: Critical Path Initiative Workshop Summary - C. Clinical Trial Design - Mary Foulkes, PhD, OBE, FDA

DR. FOULKES: Thank you, Dr. Allen.

I would like to summary the critical path thinking in terms of the quantitative sciences, in particular statistics. And you have already heard about the critical path workshop that was held last October.

I quote here from Pasteur, "Chance favors the prepared mind." This is one way to think about the entire critical path concept. And any of the scientific disciplines need to -- we need to have these prepared minds to make any advances in any of the scientific disciplines.

Another reason that this particular quote caught my eye is that chance is the title of a regular publication of the American Statistical Society.

In greatly simplifying the usual development of new statistical applications, and greatly oversimplifying, the usual approach is to start out by developing a new theory or model or method, and then to search around for appropriate application for that novel method.

What we suggest is in the critical path approach, that that sort of usual method be sort of turned on its end, and that in the situation where there is prior statistical approach or an existing approximation, to develop those new and improved approaches that are needed to advance the process of product development.

In terms of the potentials for critical path development in the quantitative methods, there are opportunities to, as you have already heard a little bit this morning, to maximize efficiency, but at the same time maintaining the reliability that we all need to have. Those opportunities include approving improved analytic approaches, and also as I will talk about in a little more detail, flexible study designs.

And as was already mentioned, we need to insure continued transparency and develop new best practices. Again, as I will talk about in a minute, the underlying assumptions in everything that we do need to come into play here, particularly in the statistical concepts.

The various CBER products that we deal with are listed here, and obviously in your consideration of blood and blood products, we need to take care that we are dealing with lifesaving therapies, but they also have the potential to transmit serious disease, and that needs to come into play in the evaluation of products.

You have heard a little bit this morning about the potential for rethinking or looking again at potential study endpoints, and these are all of the possible potential areas of investigation in terms of the contributions that statistics can make to the evaluation of these products, potentially looking at additional endpoint, intermediate endpoints, composite endpoints, and various assessment scales to improve the product evaluation process.

And I think Kathy mentioned a little bit earlier this morning that the manufacturing process is also very, very important. And there are numerous opportunities for improvement of statistical processes in the quality control of blood collection, and in many, many other areas that CBER deals with.

In terms of our opportunities in flexible study design, there are opportunities for improvement of the usual designs, and the usual methods of analyses. As we speed up the product development process, we will find that the parameters that we start out with early in the process -- and you saw that flow twice already this morning -- very early in that process the initial assumptions that we make might prove to be inaccurate. And there may need to be some modification of the designs that were based on those assumptions, and that is concept of flexible study designs.

There exist traditional approaches which include the usual group sequential designs, but we need to take care that the changes that are induced in a flexible design in this context are not influenced by the knowledge of interim data. And you have all seen that in the context of interim analyses of blinded clinical trials, where those that are making the interim analysis evaluations have knowledge of the interim data, but no one else does.

Now, there are emerging approaches for flexible study design that allow some flexibility in this context, for example, adaptive designs, Bayesian designs, that sort of thing. So, those are the kinds of new design considerations that we suggest should be a part of the critical path.

And the kinds of flexible designs that are under consideration for biologics, again, we may initially have less information on the kinds of design parameters that we might like at the beginning of product development. And we always have safety concerns throughout the process of product development. So, what we are looking for are optimal designs that balance all of these considerations.

Here are some examples of clinical trial design and analysis that come into play in non-inferiority trials, for example. We apply the general principles of the International Conference on Harmonization, but they present basically general principles, and don't get into a great level of specificity. And so, what we are looking for is new designs that will allow us to establish acceptable differences in a non-inferiority context. So, there are opportunities here for new developments in terms of design of trials.

And then in the context of missing data, there really are no preferred analytic approaches. And this is an area of statistical development that has really escalated in the past decade or so with the availability of high speed computing capabilities. Any approach that you take in terms of analyses of missing data involves assumptions. You cannot address missing data without making some assumptions. They may be very simplistic assumptions, or they may be very complex assumptions, but they have implications for the analyses of the data.

So, there are opportunities here for, for example, multiple imputation, and that sort of thing that may advance our abilities to analyze data, even in the context of missing data.

So, in terms of the critical path, the quantitative sciences in general have many opportunities to contribute to the advancement of product development. And I would like to suggest that we have an opportunity to gain from evaluation of new analytic approaches in this context. And that we should all consider statisticians, epidemiologists, and risk evaluators as well, as contributing to the critical path in this context.

Thank you. Any questions?

DR. ALLEN: Thank you, Dr. Foulkes.

Why don't we throw it open at this point. Any clarifications for Dr. Foulkes or for general discussion?

Let me start off with a general question. Dr. Carbone, one of the points you made was that there was the opportunity to reduce cost by focusing resources on the best candidate products early in development. In other words, you are proposing I think to sort of identify early where you should best put resources.

If there is a manufacturer that has come in with a new process or a new idea that is not deemed to be in that best candidate category, how do you gracefully tell them sorry, we are putting you towards the bottom of the stack? And are you going to get agreement with that analysis?

DR. CARBONE: Well, I want to be very clear actually. What we are talking about here is developing tools and knowledge. We have to, and this is our mission, to take all comers. We don't select or focus on some products and say others are unimportant. We take all products and deal with all products.

I think what we are talking about is where there are clear sticking points, that once a product enters in the regulatory pathway it is either the predictability that this is going to be a safe, effective, and manufacturable product. The information to be able to predict that, either a basic pathogenesis question, all the way to an assay is lacking.

And so, what we would like to have is some work with our stakeholders. Paul was saying industry, but really what we are talking about is all stakeholders -- academia, industry, and the patient populations, patient advocacy groups, the consumers -- to identify where we can provide the tools, so that people and groups who are working on products will have those tools to use to make those decisions. Not that we make the decision about what we work on and what we don't work on.

And in the case of very important products, where the light at the end of the tunnel isn't there, we would like to flip that light on.

DR. ALLEN: Thank you for that clarification.

Donna?

DR. DI MICHELE: I would like to actually thank and encourage all of you in this mission. I think we have been calling for this kind of an initiative for a very long time, and there are many things in the three presentations that actually have addressed the concerns.

I'll just say certainly specifically from the bleeding disorders world, that includes aspects such as setting standards in advance of product development, and alternative strategies for trials and product approval that exactly what the community has been asking for. And we are very encouraged actually to see this in your model.

I would only encourage you to be very active in publicizing this model, because there is a certain amount of discouragement and cynicism I think, and even in our community, in the bleed disorders community about this ever really moving forward. And I think that you have an opportunity to dispel that cynicism.

For instance, if in the workshop that is coming up on rare bleeding disorders, I think you should be presenting this model, and letting people know that you have it, and really going forward with it. For instance, coming into meetings like that, and actually applying this model to a specific problem in an active way. And I would hope that that is sort of in the planning stages.

DR. CARBONE: Yes, exactly. And it's a two way street really. In addition to making the model public, we need input from the community, the stakeholders to help us define what the important, critical areas are. So, going to those sorts of meetings, and holding meetings for discussion are very important.

I'm happy to act as a contact person through CBER. Just my first name dot last name at fda.hhs.gov will get you to me, and I would be happy to act as a contact, and work with our communications. If anybody would like us to come and hold a panel or present this model, or have discussion and get input, I would be delighted to.

DR. ALLEN: Other comments or questions?

Dr. Mied or Dr. Carbone, in Dr. Mied's presentation he talked about some of the opportunities that were identified for FDA and industry during the panel session. They are really exciting and really important. And I think my concern being are you going to be provided with the resources, especially personnel, in order to carry this out?

It's very, very difficult, if not impossible to take people who are already overloaded, overworked, and expect them to carry out exciting new initiatives like this. We need the resources. Are you going to get the support from up the line to do that, do you think?

DR. CARBONE: All I can really say, given government rules is we currently don't have -- this is an unfunded initiative currently. I think one of the comments I would like to make though in a nod to our important cadre of regulatory scientists and research regulatory scientists that when something is identified as critical, they will step up to the plate, such as the West Nile virus action, which was sort of an all day, all night, and all weekend activity.

But what you are saying is quite true. We have to be very focused right now, and our accomplishments can be somewhat limited, although we do work very collaboratively as much as possible to expand our ability to get these questions answers. But it is difficult. In some respects that's why prioritization of these is critical, because when resources are very tight, we have to really talk with our stakeholders and come up with a very, very clear and very carefully thought about plan.

DR. ALLEN: Other questions or comments?

Thank you very much. It's an exciting initiative, and I think we'll look forward to updates on it. And we would certainly like to be supportive and assist you in any way that we can.

We at this point will move to our first open committee discussion. Topic number one is safety of albumin revisited. We have a number of presentations, starting out with Dr. Laurence Landow from FDA giving an introduction and background.

Agenda Item: I. Safety of Albumin Revisited - A. Introduction and Background - Laurence Landow, MD, OBRR, FDA

DR. LANDOW: Back in 1979, when I was a medical intern, I elected to spend six months in the ICU. And I did this, because I always wanted to be an intensivist as soon I had entered medical school. And the ICU I selected for that six months was run by Eric Rackow. I'm sure the intensivists in the audience know who that is.

And there was a theory at that time that patients who were edematous had colloid osmotic pressures that were too low for their disease. And so, the goal was to administer bottle after bottle after bottle of albumin, 25 percent albumin, not 5 percent albumin, until we had raised their colloid osmotic pressure to a sufficient level that the fluid would be sucked from the interstitial space, into their blood vessels, and they could excrete that fluid, and they would get better and leave the hospital in great shape.

And so, I can remember administering 5-10 bottle of 25 percent albumin to a patient on a given day. And I once walked into the store room of the hospital, and from floor to ceiling were boxes of 25 percent albumin. That was in 1979.

Now, that aggressive treatment theory persisted at least with that group, for some time. I don't think in the general ICU community it persisted to that extent, but there were still people who believed that albumin was very beneficial. Well, you can imagine how stunned we were when in 1998, this study appeared in the British Medical Journal by the Cochrane Injuries Group.

And what they did was they did a systematic review of all randomized controlled trials published by March of 1998 that compared administration of albumin with no administration of albumin or with crystalloid in ICU with hypovolemia, burns, or hypoalbuminemia. And the endpoint was all cause mortality.

There was a big surprise here obviously, because they found if you look down below, that rather than albumin being beneficial, there was a strong suggestion, at least in their meta-analysis that it increased mortality, with an aggregate relative risk of 1.6 and as high as 2.23.

And they have also looked at looked other subgroups, hypovolemia, consistent; burns, consistent; hypoproteinemia, consistent, all greater than 1, suggesting trouble.

So, the next month in August 1998, sent a letter to health care providers. It said it is FDA's current view that the British Medical Journal meta-analysis warrants serious consideration. We encourage additional controlled trials on the use of albumin. And until the results of further studies are available, the FDA urges treating physicians to exercise discretion in use of albumin based on their on assessment of these data.

Well, a couple of years later, Wilkes and Navickis published their own meta-analysis. They had some problems with the Cochrane. They didn't think that the quality of the trials that the Cochrane had included were up to snuff. They had other comments to make.

And they did a very similar type of search using all cause mortality. And they found no effect on mortality. And they said that this finding supports the safety of albumin. Now, it's important to note that they were supported by industry. I think they would claim that industry had not role in the results and the design, and so forth. It's probably true, but I just want to bring that point out to you.

And the aggregate relative risk was 1.11 versus the Cochrane which was much great, 1.68. Similarly, hypovolemia, 1.59 versus 2.40; burns, 1.76 versus 2.40; and 1.59 for hypoproteinemia versus 1.69. All not significant, because they crossed 1, but still point estimates that were persuasive you might say, at least from a meta-analysis point of view.

Well, very interestingly, Deborah Cook and Guyatt published an editorial that accompanied this meta-analysis, and what they said was this. Wilkes and Navickis conclude that their findings should serve to allay concerns regarding the safety of albumin. But these results are reassuring only insofar as they failed to show a statistically significant increase in mortality.

In each case -- and here is the important part -- the point estimate, the best estimate of the true effect of treatment shows an increase in the relative risk of death of more than 10 percent overall. Confidence intervals estimate the range within which the true effect plausibly lies. These confidence intervals indicate a relative overall increase in mortality of 28 percent. And they get that from the upper boundary of the aggregate relative risk in Wilkes' own article.

They conclude point estimates that suggest harm, and confidence intervals that include important increases in mortality cannot allay concerns about the potentially harmful effects of albumin.

Well, finally, we have a recent study, the largest randomized trial addressing this question. The principal investigator is in the audience. This was a randomized double-blind trial comparing resuscitation with albumin or saline in a heterogeneous population of hypovolemic ICU patients.

Now, there are two important caveats. The first is that subjects with burns, cardiac surgery, or liver transplantation were excluded. So, we can't draw any conclusions about this subgroups. And there was stratified randomization at baseline for trauma, and not the trauma, not trauma with head trauma, with traumatic brain injury, but just trauma, severe sepsis and ARDS.

Again, the primary endpoint was all cause mortality. But look at the aggregate relative, 0.99, compared to a meta-analysis of 1.68, and another of 1.11. And they concluded that use of either albumin or normal saline results in similar outcomes in 28 days.

Now, they also looked at these pre-specified subgroups, and this is the disturbing one, the trauma. Just look for a second at several sepsis. Albumin appears to be beneficial compared to crystalloid. And they same thing, maybe beneficial to neutral with ARDS. But the trauma was the worrisome thing.

So, as they say in their article quite clearly, this is an exploratory analysis, because what drove this 1.36 greater relative risk in trauma was the fact that those patients who had trauma plus trauma brain injury had a very high relative risk dying. And as I understand the manuscript, and Dr. Finfer can correct me if I'm wrong, there was not stratified randomization for trauma for this group. It was just trauma, yes or no.

And finally, the industry wrote a letter to FDA last December, and they commented that these SAFE study data proved clinically that albumin is a safe therapy, and clearly refuted the findings of a meta-analysis from the Cochrane Collaboration, which was the subject of a 1998 Dr. Doctor letter from the FDA.

A subsequent meta-analysis that I just spoke about earlier could not replicate the finding of excess albumin-associated mortality reported by the Cochrane investigators. In fact, the results of the SAFE trial corroborate the conclusion of the second and more rigorous meta-analysis.

And so, the question for the committee is have data from the SAFE study resolved the safety concerns that were raised in the meta-analysis by the Cochrane group for: (a) critically ill patients in general; and (b) subgroups of critically ill patients with burns, hypovolemia, or hypoproteinemia?

Thank you.

DR. ALLEN: Any clarifying questions for Dr. Landow?

Can I ask you one quick question? In terms of practice where the SAFE study was conducted, are the products the same as are currently available in the United States today? And are other differences in medical practice that should be considered as we listen to the presentations?

DR. LANDOW: The albumin used in that study was 4 percent albumin. That study was conducted in Australia. The albumin used in this country is either 5 percent or rarely 25 percent. So, there is a slight difference in that respect. In terms of standard of care, I will let Dr. Finfer answer that, but I don't believe so after reviewing the article.

DR. ALLEN: All right, any other questions before we move on to the presentations?

Our first speaker is Dr. Paul Hebert, Vice Chair of Research, Ottawa Health Research Institute in Ontario, Canada. He will give us a review of the Cochrane report from 1998.

Thank you.

Agenda Item: I. Safety of Albumin Revisited - B. Review of the Cochrane Report - Paul Hebert, MD, Vice Chair of Research, Ottawa Health Research Institute, Ontario, Canada

DR. HEBERT: While I'm waiting for my slides, first I want to thank Larry and the FDA advisory panels that have decided to invite me. I should clarify that as usual these days, my perspective on this talk is going to be that of an intensivist who actually uses colloidal resuscitation fluids and albumin on a rather regular basis. So, I use it in practice, but I need to clarify that I'm certainly not an albumin expert. That actually is Simon.

I guess I was just joking with Simon at breakfast this morning. I'm still trying to figure out why I was invited to give the talk. And I am a transfusion person. And those of you who do critical care, like Len here and perhaps others, I did conduct a transfusion trial, and I guess as best I can figure, albumin is in blood, as we were joking this morning. So, perhaps that's why I'm here.

My claim to fame is a study on transfusion triggers. I have done leukoreduction work, and I am currently conducting a number of studies on alternatives. So, that's my perspective. As far as I can tell, I have no conflict of interest, because I don't research the topic area, and I don't receive funds related to albumin use or non-use.

I'm going to just going to undertake a very brief review of some stuff that you guys already know about just to highlight the issues. I got a nice picture off the Internet, if anybody is interested. I was going to say it's off the Harvard Lab site.

I'm just going to highlight albumin, a very common human plasma protein. The stuff we have is manufactured from blood, and you guys know more about that than I do. It has a reasonably high molecular weight. It's synthesized in the liver. It is negatively charged. We don't know that well how it's metabolized and how we get rid of it, but certainly plasma albumin concentrations can change with disease states, and that rather rapidly. Simon may talk a little bit about that.

Its primary function is to be responsible for what we call oncotic pressure. It's a very large molecule that stays in plasma, and in the extracellular space primarily, and doesn't move around from compartment very easily. And because of that, it actually has major effects on water and pressure. It binds drugs and other substances as perhaps one of its other main effects in blood and human health, and may act as a free radical scavenger.

As you heard, it is commonly used in the treatment -- or was commonly used, and perhaps less so now in the treatment of hypovolemia. It has been very well reported in the use of burns, and studied in this population, and largely because burns when they bleed or leak, seem to leak what looks like albumin protein from all of their wounds. I suspect people thought of replacing it early on in their care.

As a nutritional supplement, one of my colleagues calls this IV filet mignon as an approach to replacing or supporting TPN use. Hypoalbuminemia, just because we like to make things normal in critically ill patients; or we did. And then as a hypooncotic therapy as we heard Larry describe from Eric Rackow and others' group.

Now, this is a version of table, and you can look at the product monograph, and there is a lot of this in there. But this is taken from a standard hematology text. The only reason I highlight this slide is to say look, these are all the indications. Most of these have not been studied. There is a second page to that as well, and here it is.

So, there are a ton of indications or potential indications or reasons for use in practice. Very few of these have been the subject of proper randomized controlled clinical trials. That's actually highlighted in some of the systematic reviews that Larry pointed you to.

So, in terms of background, it's a product that was or is commonly used, as are other colloidal solutions. It is used in resuscitation of very sick people, which is where I use it most. And that is where the background stops here. I want to highlight a few basic facts, some of which Larry has already highlighted, and just a few quick reminders.

The extracellular compartment here -- these are just simple compartments of where body fluids lie. Plasma makes up a small portion of that. Interstitial space and the combination of the two make up about 40 percent of body water. And 60 percent is made up of the intracellular space.

If you administer fluids, and here this is generally what happens when you administer saline, a few compartments expand, and one does not. So, a rapid administration of saline typically will expand the plasma compartment and the interstitial compartment.

But since you can get an intercellular expansion over periods of time, but roughly speaking over the course of the first hour or two you will get an expansion of these two compartments. And since water and salt leaks out of the plasma phase, you get an expansion of interstitial space. Hence, that 1 to 3 rule we keep talking about.

What happens when you give isotonic albumin? Well, this is a product that largely stays in the plasma component or the vascular space, at least for reasonably prolonged periods of time. And so, until it gets degraded or leaks out in disease states, but it has a hemodynamic effect that is measured in much longer periods than for example saline is. So, it will expand the plasma phase, and not affect the other two compartments.

Twenty-five percent albumin, now it affects the plasma phase too, but because it's hyperoncotic, this molecule doesn't cross the vascular barriers well. It can actually shrink the other two compartments. This is what Larry was talking about. One of the ways to deal with edematous state patients is to actually give this product with the hopes that it sucks in fluid from the other interstitial and intracellular compartments.

The reason for telling you is to say that these products are not equal. At least their basic, underlying biology aren't the same. And this was a point that seemingly was missed in some of the meta-analyses. So, let me just give you a little more detail. Larry nicely highlighted the Cochrane meta-analysis. I'm just going to give you just a little more detail about them.

The first one, or at least the albumin one, which is what I was asked to talk about from the Injuries Group had 30 RCTs as part of the systematic review. It included 1,400-odd patients. The indications studied were hypervolemia, burns, and hypoalbuminemia. All doses of albumin were considered, and all concentrations.

And if you look in their table of description, Table 1, that included 2.5 percent albumin, which is not in use, as best as I can tell anywhere in the world; 4 percent, that's the Australian product that Simon will talk to you about; 5 percent, which is the North American standard volume resuscitation fluid; and 25 percent, the hyperoncotic agent which we just talked about. So, multiple interventions, multiple controls.

So, the control groups where equal either. Nothing was a control in two or three of the trials; saline in many of the trials; Ringers lactate in many of the others; and then a combination of dextrose and Ringers and others. So, the control groups were not equal, and the study questions were very different.

There were no protocols of care in most of these trials. And in the Cochrane Injuries review there was a limited assessment of quality. They focused on blinding and concealment primarily. So, that's a rough description of what they did.

Here is what they found. And I think anybody looking at this can -- anybody who is here probably read this paper at some point. So, what I want to highlight, this is one indication. Everybody seems to focus on this little diamond. That's what created all of the controversy.

Meta-analyses are much more powerful than that. This little diamond, from my perspective, is not that informative. And in this case it's truly non-informative, and in fact dangerous. If you were to see this, not knowing what I have already told you, you would see a plot figure that just demonstrates that the treatment effects from the individual trials are all over the map.

What you would garner from this is: (1) all of these trials are reasonably small, without even knowing this. What you wouldn't see is these confidence intervals are very large, and some of them are extremely large. And this is largely determined by the trial size. So, that's the first thing you would gather. Without the numbers on the side you would say, hey, these are really small trials.

These little dots here typically are different sizes, and they illustrate the weight of the effect by each of these trials, and again, they are all very small dots.

Now, if you were to put Simon's study in here, the SAFE trial, the dot would be about 5 times or 10 times the size of this one. It would be a big square here, because the treatment effect would largely be determined by the weight of that one trial. So, all of this suggests heterogeneity. Without knowing anything more, you would say hey, I can't conclude anything from this. These trials are dramatically different. So, that would be the first thing.

You would say why is this difference there? So, meta-analyses are powerful tools, because they help you identify differences in research designs. So, the point estimates are where people like to focus, but in truth what you want to know is what happened, just like you do most research.

So, this little diamond thing here, this 1.46 is what created all the news. And unfortunately, the authors helped that process, and we can discuss that briefly. But effectively, this what created all the controversy, when in truth there was a lot more information to be garnered from the Forest(?) plot here.

Here is the second version of that. This is a little more interesting. In the burn story here, the treatment effect is overwhelming. It's 2.4. If this was truthful, that's more than a doubling in mortality based on the administration of albumin. So, if this is the truth, this is a problem.

But again, you see two studies on one side, and one on the other. The magnitude of the effect is dominated by these two. Here on hypoalbuminemia, the nice part about this is, this is what you want to see. I'm just going to highlight the good part here. The good part is all of these point estimates are on the same side of the line. This suggests that most of these studies were reasonably coherent. And the treatment effect of course is on the same side.

Now, if I were a regulator, what I would do and use this information to look at and say, hey, listen, there is a potential problem here. This is hypothesis-generating. In other words, what I gather from this is one, just from this information alone, without what I'm going to tell you next is that one, there might be a source of concern here that we should further investigate.

And that's all I would say, especially if you are looking at this data -- let's say you use this approach, and I would suggest that that's a very powerful tool, especially as a cumulative meta-analysis technique, which is not what this is. Let's say a new drug makes it on the market, and then 50 trials later as the drugs gets used by investigators and everything, you start to see these treatment effects. That's a pretty potent tool to try to decide whether or not you should look at it further.

For example, you could have used this with Vioxx and might have found interesting things. So, just as an example and as a tool, this is important. But this is nothing more than hypothesis-generating, and this is why.

So, I'll going to tell you about the colloid stuff first, and then conclude from both of them. So, the types of trials. Well, here in the colloid study, because this same group -- Ian Roberts again was director and I think is still the director of the Cochrane Injuries Group. And this is another version of the same study, exclude they focused on colloids and crystalloids.

There were 37 RCTs in this, 1,600 patients. It excluded 11 RCTs, because they were too weak, and they had cross-overs and a bunch of other problems. Mortality was only available on 1,300 of the 1,600, and that is in 19 trials. So, you can see that they dumped all the small trials. All critically ill patients were included, and this included trauma, surgery, burns.

The interventions -- this is cool again, much like the same problem from before with the albumin story, but much worse. So, tons of different colloids. Albumin -- all kinds, pentaspan, so that's a hydroxyheta starch product, Dextran-70. These products are nowhere near the same. They have different oncotic effects. They have different hemodynamic effects. They have different intravascular dwell times, so they have different profiles. And they have different elimination half lives.

So, in critical care, especially in the European context where they use most of these products, where we don't, there is a science to all these things. You go to European meetings and you hear about the benefits and risks of all of these products. There are by no means considered the same by practicing intensivists.

And so, here again, I highlight some of the issues. And again, the control groups are all over the map, highlighting the fact that they are addressing different study questions.

Both of the studies, and I didn't highlight this from before, used a statistical technique in the quantitative side called the Fixed Effects Model. That's a less conservative approach to the analysis, because it doesn't address the interstudy variations in variability. And so, I would have argued not to use that particular statistical approach in these studies. And again, there are limited assessments of quality.

And this is what they found. Again, estimates are all over the place. Here again, all over the place. The diamond here is on the right side of the curve. It favors colloids. Everywhere it's on the other. Again, they go all over the place, highlighting the problem of heterogeneity again.

So, what did the authors say? And this is part of the problem. And I should highlight they still say this. "No evidence supporting albumin administration reduces mortality. It should not be used outside the context of rigorously conducted RCTs." Here was the problem, right there. The editorial said the same thing. Ian Roberts went around international news, and saying this all over the place, that this product kills people, as most of you know, hence the problem.

And then he said the same thing roughly with the resuscitation fluids with crystalloids. The inferences were supported by the BMJ editorials, and as most of you know, in the follow-up letters, a lot of this was downsized somewhat.

What were the problems? So, cardinal sin of science, way over-interpretation of your data. They were not very conservative in their assessment, and I'm not sure if that served anybody well, except for Simon. It justified the need for a large, randomized trial. So, that was the one good thing that came of all this.

Primary studies were exceptionally weak, neither concealed nor blinded, and that's just one of the problems. In a systematic review undertaken by Deborah Cook and Peter Choi they highlight a number of other difficulties with study quality. These authors just focused on two, but they were weak statistical heterogeneity, a huge problem. Use of fixed effects model, again a less conservative test of aggregate treatment effects.

They combined different interventions, again, a problem with homogeneity or heterogeneity, but this is of the clinical nature. Clinical heterogeneity here is a huge problem. They included in many of these things, neonates and adults, different indications, different control groups, no protocols of care. And the trials spanned two decades. And I'm just highlighting the big points here.

The other big problem was the mortality data was actually pushed by just a few trials. So, in these trials there was significant imbalances in baseline characteristics, because they were small. And that's just the short laundry list of issues.

Now, there were a whole lot of other systematic reviews, and Larry highlighted that fact for us. He highlighted the Wilkes one. They were relatively consistent in their finding of treatment effects, because they were looking at the same studies. But they interpreted the results in much different ways.

We heard the Wilkes difference. That was not found here. It's still the same side of the line, as Larry highlighted. This is the combined treatment effect for albumin. And you can see that they are on the side favoring crystalloid SAFE 1, and that's the older one. There is not much to say here except they are using the same trials and have the same problems.

They do highlight some subgroup differences. Again, I'm just highlighting the fact that there were differences in possible positivity amongst all of these heterogeneous results.

So, why do meta-analyses report discordant results? I think we have just gone through all of them. This is just a list, a paper by Alex Jadad and Deborah Cook and George Browman published in a Canadian journal. There are variations of this same publication in the Annals of Internal Medicine.

But we identified population differences, differences in intervention and underlying study design. The data extraction techniques, while we heard that the different meta-analyses did that in fact, because they included different studies. They assessed quality differently. The statistical methods were different, many of them. The section criteria we talked about.

So, this particular exercise of this particular type of meta-analysis illustrates all of this. So, this is a really good example of discordant results and discordant meta-analyses.

The types of discordance we have just talked about. There were directionality issues in several of them, one. The magnitude of the effect was certainly different in all of them. The significance and interpretation of the results were dramatically different amongst the different systematic reviews.

What happened after the BMJ thing? This was actually a publication by Roberts et al., which actually Brian McClellan, who several of you know, was the second author here. I'm sorry I didn't include him, there were only two. This is time series data. And unfortunately, it's after 1998. I don't have the follow-up to see if it's gone up again. But it just highlights the fact that in Great Britain, man did albumin use drop off the face of the earth. I'm sure that happened in Canada, and it probably happened here in the United States.

This is a recent survey that we are doing, because we are doing a small study. Well, we are actually hoping to do a large trial, but much different than the one that is currently out there, the SAFE study. As part of that, we always undertake surveys of the practicing physicians to find out if there is a practice variation, and equipoise on issues.

So, this was concluded and given to my by Laura Lynn McIntyre, the lead investigator of a pentaspan study that we wanted to undertake. The denominator here is 210 Canadian intensivists. That's almost all of them, just in case you guys want to know. It's not a big community.

And what we found is just look at albumin use here, 5 and 25 percent. It's just almost non-existent. But most of us will use a colloid solution, pentaspan. What I need to highlight here is the evidence of albumin is a whole lot better than it is for pentaspan, but we use this product, because it's the one licensed for use in Canada currently. There are two or three I guess licensed for use in the US. But this particular product is the one that has replaced albumin use.

And what you can see is in the context of severe septic shock, which is what this context was about, and this was a scenario-based survey, we use a lot of colloid still. That's in Canada.

I want to highlight two studies, and then I'm going to quit. I want to highlight two studies for a few reasons. One is to indicate that albumin is not always bad, and there may be indications for use. Of that long laundry list of things, I wanted to highlight two simple studies, because this is why a lot of us use this product.

This, Larry, gets back at why you use it, or were asked to use it, at least in part. So, this is a study undertaken by Polly Gavars(?) and Greg Martin from Atlanta, from Emory -- no, sorry, he's from Vanderbilt. He works with Gordon. So, Greg works with Gordon and they undertook a nice little study. Greg did great work on this very small, randomized trial, but very well conducted with the use of 25 percent albumin added to furosemide.

Now, the intent here was simply to demonstrate whether you could change pulmonary physiology by making people urinate and getting rid of volume, a very simple little question, nicely designed. There were 37 patients, 17 and 19 per group. And they ran this for 5 days. So, there is more work yet to come on this very question.

What did they find? So, this small study, as I told you, very small numbers, they noted a huge difference in weight loss in the albumin treated group. And again, remember, both groups were edemonous to start with. Both had ALI, and effectively both were getting lasics infusions to get rid of volume.

One group was supplemented with albumin 3 times a day, 100 mLs of the 25 percent solution. And one group loses a lot more weight. This suggests that that suck in drive thing that we do, that rinse and spin thing actually maybe works. They improved oxygenation. Now, what I didn't tell you is that that was largely an early improvement and was less true late.

They improved potentially hemodynamics. That means a decrease in heart rate, and an increase in mean arterial pressure. It's probably clinically important. But no change in other pulmonary lung mechanics. That may because you get rid of water, which is only part of the problem, and you're left with all of this underlying inflammation in the lungs that you can't change easily. I don't know. I'm not sure how you interpret that. Len's more of an expert on this topic than I am.

But this particular study is why intensivists still use albumin. I go around in the unit, and if I have a really difficult patient to ventilate and he's a edemonous, I certainly think of this.

Now, here is another one. And I'm just highlighting two positive things. This is what I'm not supposed to do, right? But whatever, because I'm cherry picking from the literature. But this is another nice and interesting study that was well conducted, conducted out of Spain. The question they asked was does albumin supplementation improve outcomes in patients who have cirrhosis and spontaneous bacterial peritonitis?

The concept here is that these patients get septic, their intravascular volume drops. And as part of that, their effective circulating volume also drops, and they go into renal failure and die. That's the simple thing. So, they vasodilate even more than they were previously, blood pressure drops, kidney function drops, and they go onto death. That was the underlying premise for this trial.

So, they are thinking, well, if we replace them with a lot of albumin while we do a peritoneal syntisis(?) and drain them of that stuff, they will do better. It seems like a reasonable question. It was multi-center trial, 125 patients. I should comment that this is a relatively rare disease, hence the small sample size. And they used 20 percent albumin, not the product that we use, but close enough.

And then the intervention was an antibiotic or antibiotic plus albumin infusions. Now, this is an infusion of albumin over a few days I believe. And what they find is resolution of infection, no different. These are all well defined in the paper. Renal impairment was a 50 percent decrease in renal function or dialysis dependence, so a sensible, rigorous definition. Mortality, you can't argue with that. And some of these were actually decreased, interestingly enough with albumin use.

So, both the studies I have just presented highlight why physicians continue to use albumin, despite the fact that they use it a whole lot less than they used to. My suspicion is though that it's creeping up again, again, non-scientific. And that there potential benefits to the product despite what we have already heard. It's unclear yet what that benefit truly is.

One other issue I'm going to highlight is do protocols of care and timing matter? This is recent data. I would just like to highlight it. Well, it's not that recent anymore. It's a study by Manny Rivers. It kind of put a whole lot of us on our ears, despite the fact it was a single center and small study. It essentially justified the existence of intensivists in many ways, and certainly fit what we believe to be our paradigm of care.

So, what he did is undertake a study of early aggressive resuscitation, including albumin, because Manny uses a fair bit of colloid and albumin in his resuscitation strategies, which is one of the reasons I highlight this paper. And as I'll show you, he saves lives.

So, what they believe is that as people get really sick and they are septic, their blood pressure drops, and the ability to profuse vital organs is compromised, because your blood pressure is low, because your microcirculation is abnormal. And this is a time-dependent phenomena.

You get sick, sick, sick, sick, sick. And then you get ischemic and the vital organs, and at a certain point if you don't fix this, you die. So, you need to fix this quickly, and then you can recover. If you don't, you go onto organ failure and an oxygen debt, and then death. And this is a time-dependent phenomena, just like stroke or myocardial infractions are.

To justify his tenet, he basically said there are a lot of studies that looked at early aggressive resuscitation, hyperdynamic resuscitations, super normal values. And when they applied early and late, they do different things. So, late application studies don't do anything. Early application studies, when they are done in early phases of shock, seem to have benefit. That was the rationale for what he did.

What did he do? He wanted to determine whether early goal directed therapy targeting treatment with venus hypoxia, did that improve outcomes? There 263 patients randomized. They were blinded assessments after the intervention, but it was an open label trial. And then he looked at goal directed care versus not.

Fluid administration was dramatically different by group, almost a liter and a half difference within six hours only between the two groups. So, that's a huge effect. It's a six hour intervention, and more transfusions too. This is where I got into the act, or have gotten into the act with Manny on a few occasions. And no difference in other therapies. So, fluids and blood were different.

What did they find? A dramatic decrease in death, almost a halving of a mortality. So, that is 30.5 versus 46. So, this is the relative risk of death, and you see the confidence intervals here. The P value is 0.009, if you are interested. So, a significant difference in relative risk reductions, a 34 percent there I have. I had to calculate a lot of these things, because they weren't in the paper. And the number need to treat is down there, it's 6.

So, what can we infer from all of this? And this is where I conclude. First, the type, timing, and quantity of fluid resuscitation may impact on mortality. That was one of the things I tried to say here. It's a complex area of care, with very, very few high quality studies. Meta-analyses primarily highlight deficiencies in the literature, and give very little in terms of guidance for treatment options.

Can't and should not infer treatment choices based on most of these meta-analyses. Albumin may be of benefit in improving oxygenation, supporting patients with cirrhosis. This is a different indication than Simon's studies. It just highlights the issue that we need more studies. And early aggressive fluid resuscitation may save lives. That's the Manny Rivers thing. And colloids may be a part of that.

I should also highlight that we have now with SAFE, we have more evidence for albumin than anything else, and most of us use something else, which I find very interesting.

Anyway, finally, what do we need to do different? We need to do studies that look at different concentrations of albumin. We now have data from SAFE, which you are going to hear about. We probably need to explore crystalloid resuscitation. There is new literature suggesting that not all crystalloids are created equal either.

Different colloids also need to be studied. Again, there is virtually no literature on a lot of these things, as you saw. All crystalloids aren't created equal, as I just said, and neither are all colloids. And then this is a hot topic for a lot of us, what in treatment protocols actually helps, and do that?

Finally, we're going to talk about SAFE, and that's where Simon comes in, and he's going to tell you all about his study.

Thank you very much.

DR. ALLEN: Thank you very much. That was a very nice presentation. Questions or comments?

DR. KUEHNERT: Thanks for the presentation. I just had a question about the SBP paper that you presented on the what did they find slide. It looked like that there was a higher hospital mortality, renal impairment, hospital mortality and three month mortality compared to control with albumin versus control.

DR. HEBERT: If that's what it said, it was a mistake. It's the other way around. Did I get the tables wrong? I'm sorry about that. Then the treatment effects were on the wrong side, and I apologize for that. The differences were all in favor of albumin, not the other way around.

DR. KUEHNERT: Okay, and the other question I had is there were some data from the SAFE study about using albumin in the setting of severe sepsis. And I wondered if you could comment on that?

DR. HEBERT: If you wouldn't mind, I would like Simon -- he is the expert, and he's in the room, and he will tell you what he believes about that. And if you wish, we can talk about my views on it after. I would like Simon to basically say what he needs to say about the SAFE study. We've got the guy who did it in the room.

DR. KUEHNERT: Sure, I just meant in general.

DR. HEBERT: I think the SAFE study is the best of its kind currently. It is a very large effectiveness type study, one that we need to replicate in other disease state with these products. Personally, I think the subgroup analyses as part of the SAFE trial are exactly that. They are subgroup analyses. They were reported by the authors, because they were a priori subgroup analyses, which is important. This is what they planned to do. And they clearly indicated in their paper that there are really no differences here, but these are just speculation of things that you might look for in the future.

So, I certainly wouldn't over-interpret the subgroup. I think that is highlights an issue for further study. And I would say the same for trauma, by the way. The inferences from trauma you might say favor harm, but when you dig in -- the trial was neutral, right? It was negative. So, that's the message you walk away with. If you start salami sciencing the thing, you are going to find differences by virtue of the fact that it's a zero sum game.

So, if you find a difference in one subgroup, by definition you are going to find harm in another. That's just the way the game works. So, that's the one time when subgroup analyses are powerful and inappropriate. In this instance the authors did it as an explanatory analyses, which Simon will highlight, I'm sure, and he will say what he believes from it. But I think he has been very cautious in how he has played that up.

DR. ALLEN: Okay, thank you. Dr. Doppelt.

DR. DOPPELT: I understand the rationale initially as to why albumin might be beneficial. But if you believe the data that it doesn't help, by what mechanism would you postulate that it is doing harm?

DR. HEBERT: Well, let's be clear, I think at this current time, the data suggesting benefit or harm is neutral. The SAFE effectively says that it doesn't change mortality. Does that mean there are no benefit beyond mortality? Yes, there may be. Perhaps we don't change 28 day mortality, but we reduce ICU stay for example in some very small subgroups. That's just a plausible explanation, one.

Two, the only drug that has been well studied is 4 percent albumin. Twenty-five percent albumin is a different product. It's markedly different. Its oncotic effects are dramatic. You are comparing it to an iso-oncotic product in this trial. So, if you are an intensivist, you will extrapolate in different ways, like we all do in clinical practice. And I'm not sure that's right or wrong, but the data supporting either approach is not there.

What Simon will highlight I assume is that overall, even if you use the product, you don't kill people. There are two interpretations. One is you don't benefit. The other is you don't hurt them either. So, if you have a specific reason to use the drug beyond what is stated, like mortality, I think you are not going to hurt them.

DR. DI MICHELE: Greetings from a fellow Canadian, and thank you for your presentation.

I was just wondering if you could clarify something for me. In not being able to actually review all the individual papers on which these meta-analyses are based, I had a hard time extrapolating from the papers the exactly definition of hypovolemia, of hypoalbuminemia, hypoproteinemia.

And to a certain extent also, I mean when you include burns in these categories, and these were very well defined categories it appears, except that they weren't well defined to someone like me, what percent burns are included in these kinds of studies? In other words, what kinds of patients are we talking about? Because this committee is being asked is albumin safe, and is albumin safe in every group of patients? But I'm not exactly sure that I can define them.

In the SAFE study there appeared to be certainly CVP. There was some analysis of CVP and albumin in both groups. But I still don't know how these groups are actually defined, and maybe as an intensivist you can help me out with that.

DR. HEBERT: I can't help you. Your observations are very astute. I think the bottom line is you just highlighted another deficiency in most of the papers. Most of the papers did not have clear definitions for their subgroups, one. Two, they are somewhat artificial. Most critically ill patients are by definition, hypoalbuminanemic.

And in fact, I think Simon was telling me this morning at breakfast 40 percent of his patients could have been reclassified in a different subgroup. He has been classified in the hypovolemic subgroup, but most critically ill patients who are hypovolemic are also hypoalbuminanemic. I'm not sure why that is, but that's probably true. So, you're asking did they define their patient groups well? The answer is no, except for very broad definitions.

And you also highlight the issue of treatment protocols. Again, I thought I did that a bit, but the treatment protocols weren't clearly defined in a lot of them. In fact, most of them did not have hemodynamic endpoints in their trials that were very clear. And the reason for highlighting Manny River's study was that when you do look for hemodynamic targets, and you are early and aggressive, you do seemingly detect differences in mortality. But the trials didn't do that.

The point of that slide and that part of the presentation was to highlight if you do it well in a single center, you may find differences. But the 30 patient studies were done two decades ago, or were very old, and didn't do a lot of those things. So, not finding differences in those studies is irrelevant today, I think. I'm trying to be nice about it, but the fact is Ian Roberts significantly overinterpreted his data, and the proper cautions weren't applied. And that is how I would review that.

DR. ALLEN: Okay, we need to move on. Dr. Hudson.

DR. HUDSON: Paul, very nice review, thank you.

I was concerned in reading the Cochrane analysis by using death as the endpoint, when it wasn't reported in a lot of studies, and like a good meta-analysis they went back and found that. And they list this as a limitation, but I wondered if you knew anything more about it, because what they said was they asked for death in the specified period of the study, which in a lot of the cases was either unspecified or was one to five days.

And then they report that the data that they got on death were either at a weak or the ICU stay. Which even though they might be the same in study, the weak and the control and the treatment group, that isn't very long in small numbers. And in critical care we know that can change enormously from 28, and maybe even to 60 days. So, I wondered about your thoughts about that.

DR. HEBERT: Well, I think you are just highlighting another deficiency. Dr. Roberts and his colleagues actually make that point in the colloid paper in a sense, where they say we were not able to identify it in 20 percent of our sample. It's not quite that much. It's 1,600 down to 1,300 roughly. They say we weren't able to find it.

And in all those who they did find death, I think it is also questionable. A lot of the studies did not identify death as an endpoint, and they had trouble finding it from the authors. So, it's just another deficiency of the approach. And again, the authors were zealous in their conclusions because of that.

DR. HUDSON: And I guess a second point I have is maybe asking a comment, but I would like you to comment on it. And that is in the rationale which you showed us the idea with albumin is that it stays within the intravascular space and pulls fluid in. But we know in these critically ill patients, particularly septic or injured patients it doesn't necessarily stay in.

And yet, we also don't have a good test, which I would have thought we'd have by now, any way to measure the movement of albumin or other large molecular weight substances out of the vascular space. Do you want to comment on that?

DR. HEBERT: Well, again, it's not my area of expertise. You probably know more about it than I do. When I was in the cardiovascular research laboratory, in fact in Jim Russell's place with Keith Walley(?), we tried to do measurements of lung water in people. And you know we can do it, lung and water outside the vascular space.

The difficulties are with the radiolabeled tracers, as best as I can understand. We had problems with signal to noise ratios. So, my view is that Greg Martin's stuff illustrates the fact that you can do it, despite where this stuff goes. It's a small study. I'm not sure what you believe of that.

DR. HUDSON: I think that's very provocative.

DR. ALLEN: Dr. Kuehnert, did you have another comment?

DR. KUEHNERT: That's okay, I can defer it.

DR. ALLEN: Any other pressing comments or questions? All right, I'm going to take the chairman's prerogative. We have been sitting for three hours at this point. We are going to take literally a 10 minute break, which is half of what's scheduled. We are running almost an hour behind. I think it's a very important discussion, however.

[Brief recess.]

DR. ALLEN: Our next speaker for the morning will be reviewing the SAFE study, Dr. Simon Finfer, Senior Staff Specialist in Intensive Care, University of Sydney, Australia.

Thank you very much for flying all the way to get here to make this presentation.

Agenda Item: I. Safety of Albumin Revisited - C. Review of the SAFE Study - Simon Finfer, MD, Senior Staff Specialist in Intensive Care, University of Sydney, Australia

DR. FINFER: Thank you. I would like to thank you for the opportunity to be here. It's very interesting to observe your processes. Obviously, everyone has heard of the FDA, but not many of us have the opportunity to see how you work. And I will be reporting back to my colleagues.

My wife would like not to thank you for inviting me here, as I was in Canada two weeks ago, and this was an unscheduled trip. And for anyone who is geographically challenged, this is a true map of the world. This is where we conducted the SAFE study in Australia and New Zealand. And I have had to travel down here, because this is where we are, down at the bottom of the world here. And that's only about 11,000 miles.

Just before I actually get on to my presentation, I would just like to address a couple of the issues that were raised in discussions that I can possible illuminate a little bit. First of all, the standard of care in Australia and New Zealand intensive care. Australia and New Zealand both operate systems of closed intensive care units, which means that everyone is cared for by accredited ICU specialists. There is not a system like exists in units in the US where surgeons, physicians, community physicians, et cetera, could admit someone to an intensive care unit and provide care.

In terms of does that make a difference? Well, I believe the Society of Critical Care Medicine has been promoting that model very heavily for the US, because of evidence that it actually produces better outcomes. And certainly, if you look at standardized mortality ratios for Australia and New Zealand intensive care compared to the rest of the world, you will find that they compare very favorably.

The definitions used in the Cochrane meta-analysis papers to determine whether patients being treated for hypoalbuminemia, hypovolemia, or burns -- burns is fairly straightforward, but the others were really just what the authors of the papers thought they were treating. There were no clear definitions at all.

I have just been through again, all the hypoalbuminemia papers included in those analyses, which was just 10 papers in the latest update of Cochrane, which was published in 2004, which includes the SAFE study data. And for hypoalbuminemia there were 10 trials with 86 deaths in a total of 657 patients.

Four of those trials were in neonates, which included 41 deaths and 181 patients. And the indications for giving the albumin were things like to see if it improved oxygenation. And yet, they ended up in the hypoalbuminemia group. So, I think what Paul said is absolutely right, they were poorly defined, and they were aggregated together as best the Cochrane Injuries Group could manage.

So, to move on to talk about the SAFE study, this was a collaboration -- and I apologize that some of the important pieces of data are on the bottom of my slides, and if you are not actually sitting here, you can't see the bottom of the slides. Although these people over here in these seats will have an excellent view.

This trial was a collaboration of the Australian/New Zealand Intensive Care Society Clinical Trials Group, of which I was until very recently, the chair, the Australian Red Cross Blood Service, and the George Institute for International Health at the University of Sydney.

We have already heard the background, which was basically the Cochrane meta-analysis. The response of the Australian government agencies and individual state health departments was going to be very similar to the response from the FDA and the Committee of Safety of Medicines in the UK, which was to issue an advisory. And we approached the government and said we would prefer to respond to this controversy by trying to answer the question rather than adopting that approach. And fortunately, we were able to get government support for that.

So, we conducted a multi-center, randomized, double-blind, controlled trial to compare the effects of two resuscitation fluids, namely forps(?) and albumin or normal saline on 28 day all cause mortality. And our population was ICU patients requiring intravascular volume resuscitation.

Our goal was to conduct a 7,000 patient study in 16 ICUs over an 18 month period. That gave us a 90 percent power to detect a 3 percent difference in absolute mortality from an estimated baseline mortality rate of 15 percent.

We pre-defined three subgroups: trauma, because the meta-analysis from Deborah Cook's group with the lead author of Choi suggested that colloids increased mortality in trauma patients; patients with severe sepsis, because as Len and Paul intimated, there is a belief in the critical care community that in those patients albumin leaked out of the vasculature, and that might be a group in which it would be particularly harmful; and patients with the acute respiratory distress syndrome for a similar reason.

This trial was conducted based on being a large, simple, effectiveness trial using the uncertainty principle. So, the main inclusion criteria were simply that the patient was in the ICU and needed intravascular fluid resuscitation; the treating doctor had substantial uncertainty over the best fluid to give, and really if anyone was scientifically honest with themselves, everyone should have had great uncertainty, because the evidence out there was very poor.

We obtained informed consent if possible, but we had a provision under local legislative regulations to obtain delayed consent, as this was classified as emergency research. And this provision was used in 95 percent of cases.

We excluded anyone of age less than 18 years. Patients with burns were excluded, so this study cannot help you answer the question in that population of patients. The reason for that exclusion was that we only had two burn centers in the study. Albumin formed part of the protocolized care of burn patients in both those centers. The number of patients we would enroll would not be enough to answer the question, and arguing with burn surgeons is just too much like hard work.

Cardiac surgery patients were excluded, because the event rate, i.e., dying is not high enough in cardiac surgery patients. The average mortality for patients undergoing cardiac surgery in the Australia and New Zealand, according to the ANZIC(?) database is 0.98 percent. So, the opportunity to study those patients, looking at a 28 day mortality outcome was just unrealistic.

Again, liver transplantation, we had one liver transplant unit in the study. The same reasons as burns, it was not feasible. It wouldn't provide an answer to the question in that group of patients, so we left them out.

Obviously, some people have a religious objection to receiving human blood products, and were excluded. Similarly, anyone with a suggestion of prior allergy to albumin was excluded. We didn't study patients who were clearly going to die regardless of what fluid they required, so anyone who was moribund and expected to die in the next 24 hours was also excluded.

The study was conducted using a fairly state-of-the-art Web-based system. We conducted all randomization data capture, fluid distribution systems, et cetera, over the Internet. And this is what enabled us to do this large study relatively quickly.

The study was funded predominantly by government money, mainly peer reviewed funds from the National Health and Medical Research Council in Australia, the Health Research Council of New Zealand. And then we went around with the begging bowl, cap in hand, and obtained additional funds that we needed from most of the jurisdictions in Australia and New Zealand, sometimes through individual hospitals or from state health departments.

We had one commercial sponsor, which was CSL Limited, based in Melbourne in Australia. They are the blood fractionators for Australia and New Zealand. CSL used to be the Commonwealth Serum Laboratories, but were then privatized. And they operate in North America and in Europe under the name of ZLB. CSL was phenomenally supportive of this study. They did not have any say in the design or conduct of the study. They played no role in data analysis, and played no role in any publication decisions.

The study treatments were 4 percent albumin, Albumex is the trade name, which is the albumin product which is manufactured by CSL and available in Australia and New Zealand. For various reasons to do with the blinding, we also had to convince them to manufacture the saline for the study. They don't normally manufacture saline, so this became a study using an unlicensed product, being the saline they manufactured. And again, this was a considerable commitment that they made to the study.

We undertook the somewhat onerous task of conducting this as a blinded study. The study was blinded by placing both the study treatments in the identical 500 mL standard albumin bottle. These were then placed within this masking carton, which has a yellow, opaque window on either side, which meant that the staff administering it could see the level of the fluid.

But we did a formal test, and found that nobody could identify this fluid. In fact, when we tested this with physicians, the only positive finding was that the more convinced the physician was they knew what the fluid was, the more likely they were to be wrong. This is true of many aspects of clinical practice.

And then Tutor Healthcare in Australia manufactured for us these special administration sets, which were green, with areas where the albumin could forth masked in black. And again, this was part of our formal test, so that we were able to obtain blinding in this study.

The fluid packaging and distribution is rather boring, but was actually an extremely difficult logistic exercise. We ended up packing four boxes of albumin like this, and each of these boxes was identified by a unique six digit code, which the Web site then knew where each of these boxes were, which hospital it was in, whether it had been activated or not.

And so, for each individual patient the nurse went to the Web site when they wanted to give fluid, entered the patient number, and the database told them which box number to go and get out of the store. And obviously, that box had to be in the right hospital, in the right ICU, et cetera. So, I don't believe we could have done this study without it being Web-based.

This just shows the system in the clinical setting, and often was asked by medically qualified relatives of our patients what it was. And when was I going to ask them permission to be in this study?

The study, as I said, was conducted very much as a large scale simple effectiveness trial. So, the treating clinicians decided the amount and rate of fluid to give based on the patient's clinical status, and their response to treatment, which is how we believe fluid is administered in clinical practice.

The allocated study treatment was used for all fluid resuscitation in the ICU from the time of randomization until 28 days, death, or discharge of patient. If a patient was re-admitted to the ICU within 28 days, they went back onto their study treatment. And we didn't try to control the administration of fluid outside of the ICU, because that was not practical. All other aspects of patient care were performed at the discretion of the treating clinicians.

The primary outcome was death from all causes at 28 days. Secondary outcomes included survival during the first 28 days, the proportion of patients with new organ failures, and the duration of mechanical ventilation, renal replacement therapy, ICU and hospital stay.

There were two pre-planned interim analyses conducted following the recruitment of the first one-third and two-third patients. These were reviewed by the independent data monitoring committee, which was chaired by Prof. Sir Richard Peto in Oxford, and on both occasions obviously they told us to continue the study.

We managed to recruit just over 100 patients per week to the study, so we completed the study 3 months ahead of schedule. This is a rather unusual graph, because the blue bars are the actual recruitment, and the maroon bars are the planned recruitment. And it is quite unusual to find the actual recruitment exceeding the target recruitment in such a trial.

This is a simplified consort diagram describing the flow of patients through the study. As you see, we randomized 7,000 patients. There were 2 patients who were accidently randomized twice into the study, right at the end of the study, which meant we ended up with 6,997 patients to analyze.

The patients who were randomized twice in error, as soon as the error was realized, they were allocated all treatment according to their first randomization, and their data was analyzed in the group to which they were first randomized, as analyses were conducted on intention to treat principle.

There were, as you can see, a 1 to 1 split of the patients. We lost less than 1 percent of patients to follow-up at 28 days. And the majority of those patients lost to follow-up was due to them withdrawing consent or withholding consent to the use of their health care data, which is obviously a right in both Australia and New Zealand. Even if you have completed a study, you can come back and say I don't want my data used, which causes us some problems, but that's the law and we have to live with it.

And so ended up analyzing 99.3 percent of patients in the albumin group, and 98.9 percent in the saline group. At baseline the groups were very matched for the major baseline characteristics. There was actually a slight error reported earlier. The stratification of randomization was solely by trauma. We didn't stratify by severe sepsis or ARDS, but they are a priori subgroups. And because of the size of the trial, they were very well balanced between the groups even without stratification.

Physiological variables and treatments being received at baseline were also very similar. You will note that there is a 0.4 of a millimeter of mercury difference in the central venus pressure, which doesn't sound like an awful lot, but because of the size of the study, the P value for that comparison actually comes out at 0.03. But that is without correcting for the multiple analysis. I think that's just one of those things.

In terms of the amount of study fluid administered, as Paul alluded to, we have all been taught this 3 to 1 rules, that 3 liters of crystalloid equals one liter of colloid. In this study that is not what we found. And in fact, the data for the 3 to 1 rule I think you will find is rather weak.

There was more saline administered. The maroon bars are the study fluid, so that's saline administered, and the blue bars are the albumin administered. And you can see for the first three days when the majority of fluid was administered to the patients, the P values suggest these differences did not occur by chance. The ratio of albumin to saline administered over the first four days was approximately 1 to 1.4, not 1 to 3.

Interestingly, and somewhat expectedly we found that patients who were assigned albumin actually received a significantly larger volume of packed red blood cells. It might be of interest to this committee, obviously. The difference here obviously generates an impressive P value. The mean difference per randomized patient over the first four days was 71 mLs.

We believe that was due to patients who were given albumin just got a better and quicker intravascular volume expansion, and then it tipped them over the transfusion threshold, which in Australia and New Zealand intensive care practice is determined by Paul's study, because we have studied that. And the transfusion threshold is round about 70 to 80 grams per liter.

The resuscitation endpoints within the trial, even though this was all left to treating clinicians to treat the patient as they saw fit, and then there were no strict protocols, unlike Manny Rivers' trial for instance. This is the mean arterial pressure at baseline and during the first four days. You actually can't see the circles representing saline, because they are obscured by the albumin, but the values were exactly the same for the first four days of the study.

With heart rate there was a slight difference on day one. The P value for that comparison of less than 0.001, again, determined by the large sample size, but otherwise very similar.

Central venous pressure was somewhat different, as I have told you. At baseline there was a difference of 0.4 of a millimeter of mercury. And this increased during the first two days of the study to an absolute difference between the means of 1.2, and then started returning towards normal. Again, the P values for those comparisons were very small.

We don't know what a difference in central venous pressure of 1.2 millimeters of mercury means, but blood pressure research suggests that even very small values such as that might be significant. And again, this supports the hypothesis that the albumin patients were getting a more rapid and more effective volume expansion.

We were very pleased to see that separation in serum albumin levels between the two groups, obviously with the albumin increasing in those assigned albumin, and decreasing in those assigned saline, because this told us two very important things. One, that CSL actually were putting albumin in the bottles. And secondly, that our blinding system was working, because there was no breaking of blinding during the trial. And it was very pleasing for us to see that that whole system, which was somewhat complex, had worked.

The primary outcome I think would be well known. There was no difference really in the relative risk of death between the two groups, with a relative risk as you have already been told this morning of 0.99, and 95 percent confidence intervals of 0.91 to 1.09. And the P value for that comparison being 0.87.

This is the Kaplan-Meier curve, the estimates of probability of survival over the first 28 days. As you can see, the curves are essentially the same.

There were no significant differences in the secondary outcomes, days of mechanical ventilation being very similar, days of renal replacement therapy.

The organ failure data is actually very complex, because these are sequential measures over time in multiple domains. And we chose to present this in this way in the New England Journal article which was the number of patients developing new organ failures. And as you will see, that produces almost no difference between the two groups.

There is in fact in our preliminary analysis, the suggestion of a difference in the pattern of organ failures between the two groups. And we are currently analyzing those data. And I'm certainly not going to present you with partly analyzed, possibly erroneous data. I did suggest you adjust the timing of the committee meeting, but you were already fixed into your schedules. So, I don't want to present partly analyzed data, but I'll mention a little bit more about that.

What has generated obviously a lot of interest and discussion is the subgroups. This is forest(?) plot which you can see all patients. The size of the box here is proportional to the number of deaths in each group, not the number of patients in each group. So, you will see although the size of the trauma and severe sepsis groups have numbers of patients that are very similar, the box size is different.

The things that generate interest, obviously these trauma and non-trauma groups, and the severe sepsis and without severe sepsis comparison. As Paul again pointed out, when you have a relative risk that is approximately 1 for all patients, if one goes one way, the other has got to go the other way. So, you can only claim that there is a benefit to one group if you are prepared to accept that there is a consequent detriment or harm to another group.

To look at the subgroups a little more, if we look at the trauma subgroup, the relative risk of death for albumin versus saline in those with trauma was 1.36 versus 0.96. And the most appropriate way to compare this within the context of a large clinical trial is to do a test of heterogeneity of which things like the Breslow-Bay test are probably the most well known.

Because we were looking at relative risks, we used the test for common relative risk. And the P value there is 0.04, which provides evidence that there is heterogeneity of treatment effect between those two groups.

There is some argument within the statistical community what level of probability we should take as being significant, with anything between Richard Peto at one end suggesting that it's not 0.01 or less, it's just a chance finding, to others saying once it's less than 0.1, you've got to start thinking it might be real. So, there is a considerable amount of discussion within the statistical community about what those tests mean.

What clinicians tend to do with trials like this is say well, that's all very well, but what I want to look at is the comparison of albumin versus saline in patients with trauma, not the relative risk in trauma versus no trauma. And here you will see that for the trauma group overall, the P value is 0.055 or 0.6 if you round it up, which would traditionally be considered as not good enough evidence of a difference.

You will see that there are 81 deaths in the trauma group assigned albumin, versus 59 in the trauma group assigned saline. So, that's a difference of 22 deaths. Twenty-one of those 22 deaths occurred in patients with trauma and traumatic brain injury. Now, we a priori defined this group of trauma with brain injury, because again, the mechanism of dying in patients with brain injury is very different.

The meta-analysis by Deborah Cook's group, with the lead author of Choi, was actually looking at trauma patients. I think there were five studies with 308-odd patients. And it wasn't patients with traumatic brain injury. So, this is actually the group, trauma without brain injury, that is most comparable to that meta-analysis.

As I said, the total literature prior to SAFE was around 300 patients from 5 studies. This is 700-odd, which is a bit better, but it's still very small. And these are both 6.2 percent mortality. So, with all the caveats that go with subgroup analysis within large trials, et cetera, if there is any suggestion of harm, it is in the group of patients with trauma and brain injury. Patients with trauma without brain injury, our data do not suggest harm.

We are looking at the group with traumatic brain injury in more detail. We didn't have enough data we felt, to define baseline balance in the context of traumatic brain injury. And also, 28 day all cause mortality is an entirely inappropriate outcome measure for patients with traumatic brain injury. The only reason you die with traumatic brain injury before 28 days is because you become brain dead or someone turns your ventilator off. In the majority of cases it's the latter.

And the most appropriate outcome measure for traumatic brain injury, and certainly I reviewed traumatic brain injury papers for intensive care journals, and I would not accept a paper with 28 day all cause mortality as a primary outcome. The outcome needs to be something like the extended Glasgow outcome score, a minimum of six months. Probably a year is better. And we are, therefore, doing that.

We are following all these patients. Currently, we have about a 90 percent follow-up rate, and we hope to get that a little bit higher than that by the end of this study. So, we have got vital status at 6 months and 12 months. Because it was a 15 month study, by the time we analyzed the data, we were already two years from when the first patient entered the study. So, therefore, we are doing our extended Glasgow outcome scores at two years, and that will be the appropriate outcome measure.

We will complete follow-up in June. Actually, the last patient was randomized into the study on Independence Day. And so, we are going to be completing the analysis we hope, by the end of this year.

Looking at the other group that generates interest is the sepsis subgroup. Those with severe sepsis, the relative risk of death is 0.87, compared to 1.05 for those without sepsis. The P value for the test of common relative risk is 0.06. Again, in the New England paper we said this provides limited evidence of heterogeneity of treatment effect, and it becomes very important whether you are in the statistical group that thinks 0.05 is the right cut off line versus 0.1. But 0.04 versus 0.06, we suggest that there is not that much difference there.

Again, looking at this the way the clinicians like to look at it, you can see the 185 deaths in 603 patients assigned albumin, versus 217 in the 615 assigned saline, the P value for that comparison being 0.09. Although this is a subgroup of 1,200 patients, the severe sepsis trial is actually somewhere in the top 10 of sepsis trials, because these are very hard things to do.

So, we concluded that in a heterogeneous population of ICU patients, and that obviously refers to adult patients, albumin and saline are clinically equivalent treatments, and the use of either results in similar mortality, time to death, similar use of mechanical ventilation and renal replacement therapy, a similar instance of new organ failures, and similar ICU and hospital lengths of stay.

And again as Paul suggested, we are being fairly conservative. We are trying not -- there is a great temptation to think you have done a great study, you have answered all the questions that ever need to be answered on this subject. Unfortunately, one of the things I have learned is that I always knew one study would never answer all the questions. But what I have also learned is that 7,000 patients really isn't enough.

What are we doing further? We have conducted, completed, and are about to submit for publication a paper that looks at treatment effects by baseline albumin. I'm not at liberty to provide you with great detail about that at the moment, but as presented at our clinical trials group meeting in Queensland last week, the data provide the same sort of evidence of a differential treatment effect as we saw for the sepsis patients.

So, there is the suggestion of a differential treatment effect there, but I would suggest it will again be something that we will look at and say, that is hypothesis-generating. Perhaps we need to study albumin in that subgroup a bit more.

We will write up and describe the sepsis cohort in great detail. I told you about the trauma and traumatic brain injury follow-up we are doing. We'll also be describing in another manuscript, the patterns of organ dysfunction between those assigned albumin and saline, which do on first analysis, appear to be somewhat different, although those are surrogate outcome measures.

The big question we were trying to answer was did albumin compared to saline, cause people to live or die? And that is really I think the important outcome in terms of the state of this discussion when we set out to do this trial. And while the patterns of organ dysfunction is of interest, if it doesn't actually translate through to any difference in stay in ICU, any difference in consumption of resources, any difference in living or dying, then is it important?

So, thank you for inviting me to come all this way. I'm getting on the plane to go back there this afternoon.

Thank you.

DR. ALLEN: Thank you very much. That was an elegant, and as you point out, a very large, but not large enough study. And good luck on your further analyses. I think they will be extremely important.

I'll throw the paper open for discussion.

DR. DI MICHELE: Thank you. I would concur with the chairman's comments. I have a couple of questions. The first is about the packed red blood cell usage, and the statistical difference in the albumin versus the saline group. I'm really puzzled by that. Have you had any further thoughts on why that might have occurred, or was that just random chance?

DR. FINFER: I think our collective view is that we just tipped more people over the transfusion threshold. As I said, this was done as a large, simple trial on a very limited budget. Our cash budget for this trial started off at $1.9 million Australian dollars, which at that time was not much over $1 million US, although with the current exchange rate it is $1.5 million or something. So, we were doing this on a shoestring.

We had 7,000 patients with an average length of stay of 6.5 days in the ICU. So, every data point we collected translated through to 45,000 data points. So, that's why people are surprised. We say we didn't actually collect hemoglobin or the hemoglobins that were measured. So, we can't for sure answer that question. But I think there is enough interference there from the CVP, et cetera that we just tipped more people over the transfusion threshold, and that's why people got more red blood cells. But I can't be any more scientific than that.

One of the issues about does albumin affect coagulation? Did people bleed more? And I think in the paper we said this might be a reason, but that is pure speculation. There were many, many things we wanted to do as we were doing the study, one of which was study coagulation, and we just couldn't get the funding at the time, I think mostly because people thought we were mad. They didn't think we could do a 7,000 patient study.

DR. DI MICHELE: That's amazing. Well, we're glad you proved them wrong. But it's still a little bit confusing, because you certainly think about more hemo -- unless you thought you were creating a tremendous amount of fluid shift with 4 percent albumin, which we heard in the talk before probably doesn't happen that excessively as it might with 25 percent albumin. But you would expect a little bit more hemodilution with saline. And indeed, your CVPs were slightly in the albumin group than in the saline group.

DR. FINFER: I think the saline extravates so quickly, that's the problem, and the albumin doesn't. So, the hemoglobin actually decreases in the albumin group quicker. Hemoglobin decreases in everyone in ICU. There is an obligatory about two units per week of transfusion requirement, and that is partly because people are sick and their bone marrow is suppressed et cetera, and it's partly because we are lousy at blood conservation. We take off lots of blood samples, and throw them in the bin.

And it would appear from the best guess that we have that in the saline group, the saline extravasates. The albumin group, it stays in the vasculature, the hemoglobin drops a bit. And as I have said, we have studied the transfusion thresholds in our ICUs, and they are very much determined by Paul's study.

And once the hemoglobin gets to -- if it's a relatively young, fit person, we will let it get down to maybe the high 60s. And then once it's below 60 grams per liter, sort of 69 or something, someone will finally crack and give blood. So, I think that's what was happening, but I can't be any more scientific than that.

DR. DI MICHELE: Thank you.

The other question I have is now based on this study, or how has this study influenced your practice in terms of fluid resuscitation in a heterogeneous group of patients in the ICU? And how has it influenced the practice in Australia and New Zealand?

DR. FINFER: Well, everyone asks that question. I told Paul over breakfast that's the one question I hate being asked, but it's what everybody asks. In my hospital, and I think fairly widespread in Australia and New Zealand currently we basically have prohibited the administration of albumin to patients with traumatic brain injury pending the results of that other study, because there is evidence of harm.

I don't chew people out for giving albumin to people with sepsis, but we haven't decided. We believe that is hypothesis-generating data, and we leave that to clinician's discretion.

We have a rather unusual situation in that the blood is fractionated by CSL, but owned by the government, and albumin is seen as a waste product of immunoglobulin production, and it is supplied to the hospitals free of charge. Starch is not licensed in Australia. So, in terms of colloid solutions available, we have albumin, which is free, versus other things that we have to buy.

So, that's one of the reasons why we wanted to do this study, because albumin is very widely used. I think Paul has shown you the data from Canada. I believe the data in the UK is very similar. They asked me over to talk about this at their meeting, and there was a show of hands amongst the audience of the intensive care society. They are 90 percent colloid resuscitators, and they are using a lot of starch.

And albumin actually has the best data in terms of safety, because there is nothing comparable to this for any other colloid solution at all. That doesn't mean albumin

is the better thing, it just means there is no meaningful outcome data for any other colloid solution. And for those of you who are interested, the Cochrane meta-analyses have been updated, both the albumin one in 2004, the albumin one which incorporates SAFE. And not surprisingly it is basically the SAFE study results, because it is over 90 percent of the data in the albumin in the hypovolemia group.

In the colloid one they have actually split it up by colloids this time, and all the confidence intervals are very wide, because the studies are very limited.

So, in Australia there is still quite a bit of albumin being used, although initially after the SAFE study I believe there was a reduction in the amount used. In New Zealand the individual hospitals pay by the bottle, so they are influenced by cost, and they starch available, which is cheaper. And I think financial considerations drive a lot of those decision-makings.

If you do go away and read the new Cochrane meta-analyses, I would just comment that they put the SAFE study in the hypovolemia group. And Ian Roberts is on public record as arguing that the safe study supports the results of the first meta-analysis. I'm not quite sure how you get to that conclusion, but it's basically because they say it's bad for trauma, and we said it killed people.

In the SAFE study, 40 percent of the patients had a baseline serum albumin of less than 25 grams per liter. So, you could equally well put 40 percent of the patients into the hypoalbuminemia group. And I think the Cochrane meta-analysis is not as rigorous I think as I would like it to be.

Because if you actually take the 40 percent of hypoalbuminemia patients out of SAFE, which you can argue about the validating, and dump them into the hypoalbuminemic, it pushes the treatment toward the other way very significantly.

DR. ALLEN: Very interesting comments. I was equally interested in your comments on the economics versus the scientific data that are available.

Other questions or comments?

DR. LEITMAN: I would like to ask a question. On the mean volume of fluid administered, there was significantly more saline than albumin administered. But the mean arterial pressure in the two groups was superimposable. So, I'm trying to look at reasons for which more saline was given. Is that because pre-set of threshold of hypoalbuminemia was reached in the saline group?

DR. FINFER: No, it's just because it doesn't stay in the vasculature. The endpoint is the cardiovascular -- people resuscitate with cardiovascular endpoints. They have a view that they want the blood pressure to be a certain level, the heart rate to be a certain level, et cetera, and the urine output to be a certain amount, indicated adequate intravascular volume resuscitation, and other things like respiratory variation in arterial blood pressure.

And the problem with saline or all crystalloids is obviously you pour it in, and most of it ends up not in the vasculature. So, that's why I said the rule of thumb that everyone has worked to is 3 to 1. But you would expect actually to see that to be three times as much saline as albumin, which obviously we didn't find.

DR. LEITMAN: And the reason the mean arterial pressure was identical in both groups is because it was supported with the extra saline then? Is that the conclusion?

DR. FINFER: Yes, more saline was given to end up with approximately the same intravascular volume.

DR. HUDSON: Simon, have you looked at any sort of subgroup, however you might define it, with patients that would have clinical shock or are particularly a hypotensive group?

DR. FINFER: We haven't. We defined the cardiovascular status of patients using the SOFA score, which is a thing called the Sequential Organ Failure Assessment score, developed by Jean-Louis Vincent(?) and others in Europe. And so, we have the opportunity to do that, looking at the threes and fours in the SOFA score, which indicates shocked patients.

Again, we will do that. We are very, very hesitant to sort of start pulling out post hoc subgroups, because eventually you are going to turn up something positive just by multiple analyses obviously. But we will look at it, and if we look at it and we report it, it will be very much as an hypothesis-generating thing.

I'm very interested in that, and I think Manny Rivers' paper, again where the question of whether achieving early rapid intravascular volume expansion by whatever means changes the outcome is a very important question.

DR. DI MICHELE: Just one further question. Considering albumin is a blood product or certainly a blood derivative, is there any consideration with respect to viral safety? Even though albumin has had a great track record so far, is there any consideration to viral safety in the choice of product use?

DR. FINFER: Well, in the absence of convincing data that choice of fluid affects an important patient-centered outcome, which is really the current state of knowledge, the two things that I think would drive the choice would be economics, and secondly a concern about transmission of infection.

Obviously, albumin does have an incredibly good track record, and it's way down that graph of risks in terms of the way it can be treated, et cetera, compared to other blood products. But I guess ultimately that may be a factor that affects some people. At the moment in my environment, it is probably coming way down after the economic issue, and I suspect that it is in other people as well.

MS. BAKER: Following up on Dr. DiMichele's question, is there any information about the amount of albumin that is in some of the other blood products -- clouting factor concentrates, et cetera? And is that standard by product?

DR. FINFER: That can't possibly be a question for me.

DR. ALLEN: I'm not the one to answer it either. Anybody want to address Ms. Baker's question?

DR. KATZ: The answer in my experience is yes, we know, and it's hard to imagine that it would be clinically relevant, which is different than doing the trial.

DR. KLEIN: It isn't. There is very little albumin in any of these products, although certainly you could find it if you looked hard enough.

DR. FINFER: I guess one other comment about the SAFE study is the way we designed it, was we surveyed what people did in terms of fluid resuscitation, like Paul's group. And it was from every extreme. There were people who are I only ever use crystalloid. Colloid kills people. Right up to I always use colloid, because crystalloid just leaks out and makes people edematous, and then the oxygen can't get to their cells, and they die.

And so, we chose to basically compare the two extremes, because you had to receive -- to be in this study you had to have had no fluid resuscitation prescribed and administered in the ICU, and you received the treatment, albumin or saline, for your entire time in the ICU during up to 28 days, including readmissions.

So, I think if there was a harmful effect from albumin, we would have detected it, because we were giving albumin exclusively. And that's also important when you consider is there a benefit in for instance, a subgroup of patients with severe sepsis? Because if you did take that view, it wouldn't be a question of giving the odd bottle of albumin every now and then. You would have to say any patient who comes into my ICU, who has severe sepsis, gets albumin and exclusively albumin, and nothing else for intravascular volume expansion.

So, I think if your question was concern about well, is there albumin in other things that might -- I mean, we looked to the extreme, giving considerable amounts of the stuff.

DR. EPSTEIN: Dr. Finfer, a narrow question about the blinding. You pointed out the gratifying result of a rising albumin in the albumin cohort. But were the treating physicians blinded as to the result of laboratory determinations of albumin?

DR. FINFER: No, they weren't. And in terms of protocol violations, they were minimal. The most common reason for protocol violation was error. Quite often someone who was readmitted to the intensive care unit, and the staff didn't realize they were already in the study, so they gave them some other resuscitation fluid.

Administration of other fluids, because the treating clinician felt the other fluids were indicated was very uncommon, well under 1 percent. And it was more common in the saline group than in the albumin group. But there very, very few protocol violations. So, although the clinicians weren't -- as you see, the majority of the fluid is actually given on day one, and then again on day two, by which point generally the only albumin measurement available was the baseline one.

DR. GOLDING: Can I just make one point? On one of your slides you show that there was a significantly greater use of red cells in the albumin treated group. And I would like you to comment on that, because some people would think that giving blood is a risk factor. And it has been associated with some endpoint in trials that we look at in terms of if you give less blood, it's better. Would you like to comment on that, and the risk associated with that?

DR. FINFER: Yes, it's very hard to quantify. I think if you read the discussion in the New England Journal, I just quoted from Paul's paper where the conclusion of his paper where there was some suggestion -- Paul, you'll correct me, you know the numbers much better than I do -- but the group who were randomized to the liberal transfusion strategy had a marginally higher mortality rate. Isn't that right, Paul?

But the amount of blood administered was something like two or three units more. Whereas, in our study the albumin group got on average 17 mL, which is about a third to a quarter of a unit. Now, I can't quantify is there some treatment effect from giving an extra quarter of a unit of blood to a group of patients. Well, you could argue albumin must be a little bit better, because it managed to offset this bad effect of people being killed by the blood transfusion. But this study wasn't designed to answer that, and it's getting a little convoluted.

DR. LANDOW: In the operating room it's customary to dilute packed cells with saline. What was the protocol plan for that?

DR. FINFER: We didn't prescribe how people administered blood products. I don't believe diluting packed cells is common practice in any of the units in Australia and New Zealand intensive care. It would just be administered as packed red blood cells.

DR. LEITMAN: Almost all units in the US are administered in additive solutions, so the practice of diluting with saline doesn't occur. In Australia I assume that's the same thing, that they manufactured with additive solution, which already dilutes them.

DR. KLEIN: I just wanted to make a comment that one of the risks of having a very stringent transfusion trigger is that instead of the doctor determining which patients need transfusion when, a rapid dilution can make someone fall below the transfusion trigger, and have red cells administered when probably there is no good physiologic or other reason to give them.

DR. HEBERT: Just first, I think we can't underscore how important the SAFE trial was to the critical care community. This is the first time a trial I think exceeds a few thousand patients ever. And so, Simon did us a huge favor, because several of us are now thinking of doing large trial. Len and I were talking, Simon and the Australian critical care group are now in discussions in doing trials for themselves, and in collaboration with others. And certainly in Canada and Europe we are expanding our clinical trials program, so this is a huge, huge thing.

And all of the deficiencies I identified in these Cochrane reviews, they fixed. And I think it is a huge deal. Simon, I still remember going to see you four years ago, and two years into the planning of the trial before they randomized their first patient, and how much work they went through. So, congratulations, it's amazing.

The second comment was related to -- oh, I'll just stop there.

DR. ALLEN: Other questions or comments?

DR. QUIROLO: Can you comment on the low amount of normal saline that was used? Do you think that was the randomization or some other factor?

DR. FINFER: The low amount?

DR. QUIROLO: Most people thought that there was less normal saline used for resuscitation than would be expected. I think you said that. Do you know why that was?

DR. FINFER: In terms of the ratio of saline to albumin given, it was less than people expected. I honestly thought it might be somewhere around about 2 to 1. Having looked at the literature and what is the evidence for this 3 to 1 rule, I didn't think it would be 3 to 1.

There are things that happen with clinicians. This was a very heterogeneous population. We were really looking at did albumin kill people, and the answer is no. And it's a very heterogenous population of patients. And if you have people like me, who have Scottish mothers, if you give half a bottle of something, and then you sort should I give the other half or not? And you think, well, it's open, it's hanging there and I'm going to have to throw it away in a few hours if I don't give it. And it's saline.

Well, those sorts of decisions -- and if it becomes at the lower end, and that's why I think Len and other people have been saying why don't you look at a sicker group, why don't you look at people who got more fluid? Because at the lower end, someone who got a bottle of albumin, had they been randomized, they might have got one bottle of saline, because there are so many other things happening to these patients.

They are critically ill with multi-system disease. There were people randomized in the trial who didn't get any resuscitation fluid at all, because they looked hypovolemic, and they were randomized in the trial, and someone was about to give them their resuscitation fluid. And then the resident came in and said, hey, we just the hemoglobin back. It's six. And they said, okay, well, forget the SAFE study fluid, let's get some packed cells.

So, there were all those factors that influenced it. And at the lower end of fluid administration, I think the ratio might be more like 1 to 1. And in the sicker patients, we might see the ratio separate out. But again, that's another issue that we need to look at.

DR. ALLEN: Dr. Hebert, and then we need to wrap this up.

DR. HEBERT: Multiple questions as I thinking through, and it just reminded me of a few things. Simon, one of the beauties of the trial was the fact in your first presentation I still recall how you just presented the simple, unadjusted analysis to make the point, which is what you did in large part in the article.

Have you had a chance to do any further adjusted analyses that would support what you have already said? I guess that's the first. The second kind of related question is I have you looked at in adjusted analyses, what the effect of co-interventions would be? And that goes to the issue of red cell transfusions.

And I clearly get it from the paper, but did some patients, especially in New Zealand, get other gelatins for perhaps unsteady and stuff like that? So, additional co-interventions. Blood would be one of the ones I would be interested in. Did you look at any of that stuff?

DR. FINFER: We haven't looked at that issue specifically yet. In terms of other fluids, there were so few patients who got protocol violations, and got other fluids, because if they were in the study, they only got study fluid. If they got given a gelatin for instance, or a starch, that was a protocol violation. That happened very, very rarely.

DR. HEBERT: So, it didn't happen.

DR. FINFER: No. We have just started doing some baseline serum albumin, as we discussed. But predominantly we always set out, we hoped to do a big enough trial that we wouldn't have to worry too much about adjusting for other things.

DR. ALLEN: Okay, thank you again very much, and have a safe trip back.

DR. FINFER: Thank you.

DR. ALLEN: At this point we move to an open public hearing. We have four speakers who have requested to speak. These are not necessarily dealing only with the albumin issue. Some of them are comments or points of view that go back to the opening committee updates.

Dr. Freas, do you want to make an introductory statement?

Agenda Item: Open Public Hearing

DR. FREAS: I just want to say one short thing. We have received two written comments. In addition to hearing eight oral comments, we received two written comments, which will become part of the meeting record. They are from Mr. Terry Singletary and Dr. Gary Haynes. Copies were provided. Where copies were provided, these were distributed, and where they weren't provided, they are in our viewing notebook and will be posted on the Web very shortly.

Thank you.

DR. ALLEN: I need to read an open public hearing announcement for general matters meetings. Both the Food and Drug Administration, FDA, and the public believe in a transparent process for information gathering and decision-making. To insure such transparency at the open public hearing session of the advisory committee meeting, FDA believes that it is important to understand the context of an individual's presentation.

For this reason, FDA encourages you, the open public hearing speaker, at the beginning of your written or oral statement, to advise the committee of any financial relationship that you may have with any company or any group that is likely to be impacted by the topic of this meeting. For example, the financial information may include the company's or group's payment of your travel, lodging, or other expenses in connection with your attendance at the meeting.

Likewise, FDA encourages you at the beginning of your statement to advise the committee if you do not have any such financial relationship. If you choose not to address this issue of financial relationships at the beginning of your statement, it will not preclude you from speaking.

Each speaker has six minutes, and given that we are now an hour and 15 minutes behind schedule, I am going to stick very strictly to that. When you see the red light go on, please you've got 15 seconds to summarize before I bang the gavel here. So, I really would appreciate your keeping within the time limits.

Our first speaker is Dr. Samuel Coker of Paul Medical on the Paul Smart Filter technology to remove different strains of infectious prions, including vCJD prions.

DR. CERVIA: Good afternoon. Happy St. Patrick's Day. I am not Sam Coker. Rather, I'm Dr. Joe Cervia. I'm an infectious disease internist and pediatrician, professor of medicine and pediatrics that Albert Einstein College of Medicine. And one financial disclosure is my employment for the past five months as medical director for Paul Corporation.

It's my great pleasure to address the committee this morning for just a few moments on some of the exciting work that is going on at Paul with respect to removal of infectious prion from red cell concentrates, which is a group of projects that have been spearheaded by Dr. Sam Coker, who is our principal scientists and technical director.

It's my hope that at some later time, perhaps the committee will see fit to design a time whereby we might be able to come back and discuss these data in some greater detail. Nevertheless, I'm very grateful for this opportunity.

Although neither the appearance of bovine spongiform encephalopathy or Mad Cow disease in North America, nor the transfusion transmission of vCJD, which has recently been described are entirely unexpected, these events do provide the impetus for further consideration of means to safeguard the blood supply from the risk of transfusion transmission of infectious prion.

While many feel that existing measures to deal with this risk are sufficient, there is a still a serious need to balance their potential efficacy against the safety and availability of lifesaving blood products. Animal studies have demonstrated that the agent that causes the human form of Mad Cow disease can be transmitted through blood transfusion.

With recent news confirming from the United Kingdom that a patient who received donor blood during an operation in 1997, developed variant Creutzfeldt-Jakob disease and died six years later, some transfusion medicine professionals and the general public now perceive that the risk of developing vCJD from blood transfusion may be a growing, rather than a declining threat to patient safety.

The existence of asymptomatic prion carriers raises the real concern of a second wave of prion disease. Because bovine spongiform encephalopathy has spread to other countries, and given the unknown time interval between exposures and onsets of symptoms, public health officials are faced with the complexity of estimating the true size of the vCJD epidemic.

Are there thousands of dormant carriers of vCJD at risk of developing clinical disease? This question needs to be addressed with priority and urgency, because it raises deep concerns that many more cases of undetected prion protein infection may underlie the overt epidemic, with major implications for future estimates in surveillance of vCJD.

Paul's approach to improving the safety of global blood supply grows out of our core competency in material science engineering, and a legacy of leadership in blood filtration technology. It is only natural that our solution involves a novel proprietary surface modification technology that removes all type of prions.

Because of a low tider of prion infectivity that is believed to exist in a unit of blood, the use of affinity filtration would be considered to be an efficient and effective method for reduction, while eliminating the higher costs and additional steps associated with anti-mortem detection assays, or pathogen activation processes.

Over the past 12 months we have been conducting tests to validate the removal of infectious prion from red cell concentrate using our new prion reduction filter. In these studies, which I will discuss with you in just a moment, we used all the methods that are currently available for validating prion removal. At the same time, we have completed an extensive battery of in vitro and in vivo studies demonstrating the safety and efficacy of the filter for use with packed red blood cells for transfusion.

In this one study, we initially injected hamsters with an agent that induces scrapie in hamsters. There is a particular strain that is called the 263K strain. About 200 hamsters were injected intracerebrally with the scrapie agent, and after approximately 70 days, the animals developed scrapie, and blood samples were collected into CPD anticoagulant.

About 500 mL of blood were then collected and processed into red cell concentrate using standard blood bank procedures. The red cell concentrate was then filtered with the new prion reduction filter in a manner that was similar to what I describe earlier.

You will note that the level of infectious prion in red cell concentrate before filtration was very low, and had to be concentrated in order for us to detect it on a Western blot assay. But after filtration with the prion and leukocyte reducing filter, infectious prion were removed to below the limit of detection of that assay.

In order to determine whether we had completely removed infectious prion from the red cell concentrate, we intracerebrally injected samples of unfiltered and filtered red cell concentrates into 20 normal hamsters. We then monitored the hamsters every week over a period of 300 days for clinical symptoms of prion disease, such as a wobbling gait, irritability, or head bobbing.

The hamsters that received the filtered red cell concentrate did not develop clinical signs of prion disease during the 300 day test period. In order to be certain that none of the animals had any infectious prion in their brain, we examined the brain of these animals for the presence of infectious prion.

Just as we were unable to observe any outward signs of clinical symptoms or disease, we did not find any infectious prion in the brains of the hamsters that received the filtered red cell concentrates. In contrast, 2 of the 18 control hamsters that received unfiltered red cells developed clinical signs of scrapie, while the third animal showed the presence of infectious prion in the brain at autopsy.

In order to demonstrate that the new filter is capable of removing prion that are responsible for CJD, we also used the transgenic mouse model of human CJD in which the mice were first injected with human CJD. Infectious CJD --

DR. ALLEN: Can you sum up, please?

DR. CERVIA: Okay, we'll sum up quickly. In summary, initial bioassays using the prototype filters in hamsters confirmed the removal of about four log of infectious prion. Additional bioassays using a CE marked filters are planned to confirm log removal. Our preliminary results using the scrapie model shows that prototype filters are effective in removing infectious prion from blood and blood products.

Paul Medical is completing the safety and efficacy testing required to declare a CE mark compliance on the Leukotrap Affinity Filter by spring of 2005. The new filter is expected to enter the regulatory process in the US and Canada later in this calendar year. The product represents the first of a new generation of proprietary smart filters that will concurrently reduce leukocytes and the risk of transmission of vCJD through blood transfusion.

With that, I thank you.

DR. ALLEN: Thank you. I'm sorry, but we are significantly over time.

The next speaker is Mr. Jan Bult, President of the Plasma Protein Therapeutics Association, speaking on albumin.

MR. BULT: Good morning. My name is Jan Bult. I am used to the confusion in the United States, by my name is Jan Bult, and I am the President of the Plasma Protein Therapeutics Association and a full-time employee of PPTA.

Our members provide 60 percent of the world's needs for source plasma and plasma protein therapies. These include clotting factors, immunoglobulins, apha-1 anti-trops(?) and also albumin.

We appreciate the agency placing the albumin issue for consideration by the Blood Products Advisory Committee. I had the honor of addressing the BPAC meeting last July 23, focusing on the topic of IV Ig supply. In that presentation I informed the committee about the complex set of issues regarding the economics of manufacturing plasma protein therapies, important to guarantee a stable of the needed therapies.

For a healthy industry, it is of surpassing importance to have the ability to manufacture multiple therapies from collected plasma, and for the regulations governing the industry, as well as the dear doctor letters to be science based. As one of the first therapies manufactured for collected plasma, albumin has been very important for decades.

The Cochrane publication in 1998, marked the beginning of critical comments about the use of albumin. In the United States, the US Food and Drug Administration posted a dear doctor letter detailed FDA concerns about the albumin administration after publication of the Cochrane meta-analysis.

Since then, PPTA has been actively engaged in working with experts to better understand the true value of albumin. And researchers undertook several activities ranging from basic research, to the development of other meta-analyses, but also clinical trials.

I would like to state for the record that in response to Dr. Landow's introductory comments, PPTA has supported various studies, but has never influenced the findings of the investigators. We respect the rules of ethical conduct.

The essential detail in that publication, and used as a foundation in the dear doctor letter is the suggested 6 percent increase in mortality in certain groups of critically ill patients. Many researchers have since published findings that could not confirm the conclusion drawn by the Cochrane publication, including a meta-analysis by Wilkes and Navickis that included the same 24 studies cited by the Cochrane group, as well as 18 additional studies that led to the conclusion that the Cochrane publication was based on an incomplete data set, and did not lend itself to the findings used in the dear doctor letter.

While no research paper can be considered perfect, the volume and quality of criticism leveled at the Cochrane study by neutral observers seems to be unusual. The issue regarding albumin safety does not resolve itself to the level of the Cochrane publication. First, the SAFE study.

The Vincent 2003 publication, which involved the calculation of adverse event rates for over 16 million distributed albumin doses from 1998-2000, while compensating for potential underreporting of these events, demonstrated a comparatively rare occurrence non-fatal and fatal adverse events after albumin administration. The Vincent group noted that this study adds to a large body of existing evidence indicating human albumin to be remarkably safe.

The SAFE study also adds to the preponderance of data demonstrating albumin as a trustworthy therapy. The endpoint at issue in this discussion is that of mortality. Clearly, there was a great weight of scientific evidence lending itself to the safety of albumin, and the SAFE study is an example of this.

PPTA believes that the Cochrane study has caused unnecessary concerns about the safety of albumin. PPTA supports physician decisions based upon the available scientific literature and clinical expertise. We express the hope that the committee will come to the conclusion that albumin is safe, and that the warnings given in the 1998 dear doctor letter are no longer needed.

Thank you.

DR. ALLEN: Thank you very much. Thank you for coming in under time.

Our third speaker is Dr. Gary Haynes, Medical University of South Carolina, speaking on albumin.

DR. HAYNES: Thank you very much. Good morning.

My name is Gary Haynes. I'm an anesthesiologist in Charleston, South Carolina, and a professor at the Medical University of South Carolina. Just as a way of disclosure, I have done some consulting work to a plasma products distributor, as well as pharmaceutical companies such as Nova-Nordisk. I'm the member of a speakers bureau for the American Red Cross, Baxter and Bayer as well. I guess an additional qualification is I'm also a guy who does a lot of resuscitation. I take care of a lot of critically ill patients in the OR. My practice is a full range, from pediatric to elderly.

There are two points I'd like to make today. One to add to everybody else's comments about what I think are some of the shortcomings of the Cochrane report. But also to just make sure that we are all on the same page, and that we are very clear in the statements that we make, because we have had two wonderful presentations this morning that I have learned a lot from.

And even though many of the statements that are put out in the lay press are correct, they may be taken out of context, because I have also been a residency program director, and I know a lot about how physicians learn from literature, which is often not very much. They want to just go to the bottom line and get a very salient, take away message. And if our take away message is that saline and albumin are equivalent in a heterogeneous population of ICU patients, it may be a big mistake.

So, I would just like to make the point that all ICU patients are not the same, with just a few slides. A number of meta-analyses and their value have been discussed this morning. This is one recent one that compared the use of albumin to crystalloid or various doses of albumin. And to simplify it a bit, rather than look at the -- or confuse it with looking at varying doses of albumin, just focus on the 32 randomized clinical trials where albumin was compared just to crystalloid. There was a positive value associated or favoring the use of albumin in these patients.

In another group, and Eric Rackow's work was mentioned earlier, if you look at a group of intensive care patients -- and I know I'm cherry picking here, but I'm cherry picking to make a point. And in this slide, I think the important point is to recognize that in hypovolemic patients, many of whom were septic, there is a big difference whether crystalloid was used to resuscitate patients or colloids.

Colloids were associated with about a 20 percent roughly incidence of pulmonary edema, whereas with saline it was over 80 percent, almost 90 percent developed pulmonary edema. The important message here is that the average age of the patients in this study was 80 years of age. So, we are dealing many times with a very disparate groups. We need to take that into consideration with the messages we send.

Shoemaker's work has also been mentioned. And in this fairly recent study we see where the use of albumin was just as efficacious in increasing oxygen delivery as giving a unit of packed red cells. And both were similar. Using whole blood was a little bit better, but the use of whole blood was actually associated with an increase in oxygen demand, as well as increasing oxygen delivery, but in all instances they were better using a crystalloid solution.

The Sort study has almost been mentioned this morning. I think it's a very important one and a useful one, so, I'm not going to go into the background of it, since it was very nicely covered already. But clearly, once again we see that the use of albumin in a very sick population of patients, ones that have very intense systemic inflammatory response, decreased renal failure or renal impairment, and decreased the incidence of death in a group that is very high risk for dying anyway both in-hospital and at three months.

So, based on these kinds of studies and this kind of information, there is a position statement by the American Thoracic Society, a consensus statement saying that resuscitation with iso-oncotic albumin in septic shock may be beneficial.

And as Dr. Finfer has shown from their study and the beautiful discussion he made this morning, whether albumin or saline confers a benefit in a highly selected population of critically ill patients requires further study. I'm very happy to know that they are continuing along that line.

Because given what physicians have to work with, guys like me, who are taking care of patients in the OR, in a critical situation where you don't have a lot of time to make decisions, we have to look for what we think is the most important take away message. And what physicians have to work with these days, if you take the SAFE study data of their 1,218 patients out of roughly over 7,000, you see that the decrease in mortality with albumin, the statisticians can argue whether this is significant.

I think it's quite noticeable, because when you compare it to another product that is out there on the market, activated protein C, you see a similar reduction in mortality, but obviously a very dissimilar cost associated with these products.

So, to conclude, my points are that I think the 1998 Cochrane meta-analysis had multiple flaw designs. And just to clarify, I don't know if it was specifically mentioned by either of the discussants earlier, but the deaths that came out of that meta-analysis of 24 studies were confined to three 3 reports. So, I think that in itself shows that it had many flaws.

The SAFE study tells us that albumin is not associated with an increased risk of mortality in a heterogeneous population of ICU patients. I think that's a wonderful finding. But I think it's limited only in that it doesn't demonstrate that albumin and saline are therapeutically equivalent in specific ICU patients. And so, obviously, we have to continue individualizing therapy.

The SAFE study demonstrated a strong trend indicating that albumin administration may increase survival in ICU patients with severe sepsis. I think that's something that the FDA should strongly advocate and in any way facilitate that kind of work as an extension of the SAFE protocol to enroll additional patients with severe sepsis.

Thank you very much.

DR. ALLEN: Thank you.

Our last speaker in this section is Joseph Latino, Director of Research, Lipidvirotech, regarding the inactivation of prion in biological fluids.

MR. LATINO: Thank you very much. My name is Joseph Latino. I'm Director of Research for Lipidvirotech, which is a relatively nascent biotechnology company based in Salt Lake City. I'm their director of research, and I think that fulfills my requirements for disclosure.

Lipidvirotech announced on March 3, 2005, the ability to utilize their drug delivery production system to inactivate infectious prion proteins in bovine serum. Bovine serum, as you know, has received a heightened degree of concern with regard to its utilization in biopharmaceutical products, injectable vaccines, et cetera, for the possibility of concern that infection prion material may actually contaminate bovine serum preparations, and therefore contaminate certain injectable drugs and vaccines. Anything that is developed in a hyperdome or a cell cultured medium technology I should say, will utilize bovine serum.

The company approximately a year ago embarked on looking at and developing what was known or is coined a pathogen inactivation platform. It has developed a drug delivery system and production system which utilizes -- and it sounds unique actually -- utilizes ozone and exploits oxidative potential to inactivate numbers of viruses.

It developed this platform, and this process is proprietary technology, has demonstrated a broad spectrum antiviralability to inactivate a whole host of viruses, both lipid envelopes and non-lipid envelopes. As an extension of their pathogen inactivation platform, they looked at or the literature started to support that in the isolation of native PRPSC, or native prion, there was a co-purification of approximately 1 percent of lipid associated with PRP, which includes cholesterol, single myelin, and alphahydroxycerebracide.

And although there has been no direct evidence supporting the role of lipids in prion, the fact of the matter is that prions is a GPI-anchored protein, glosamophospherial anasotile(?) anchored protein, and that anchor is responsible for at least assists in the template misfolding by PRPSC.

There was some consideration given that maybe this technology could be extended to looking at the inactivation of PRPSC. PRPSC was utilized in a cell-based assay using a murine neuroblastoma cell line which was obtained from Inpro(?). They also provided a scrapies-infected neuroblastoma cell line, which has a protease-K resistant RML strain modified for murine modified process.

Infectious prion material was purified, and was collected as an inocular from this SCN 2N cell line. And that inocular was given the benefit this drug production system, which is an ozone-based technology.

Subsequent to that, it was a proteinaceous infectious material or scrapies infectivity was measured through various immunoblotting assays as developed by actually Prusner(?) and Bosk(?), who published these papers in the Journal of Virology in 2000, where they were able to develop an N2A or neuroblastoma cell line which was not a replacement, but they say suggests has the same bioassay sensitivity as a murine cell line.

After the benefit of treating this inocula with this drug production system, which is an ozone-based system, they were not able to demonstrate either by cell blotting or by through Western blotting any detectable -- it was below the detection limits. Based upon the starting tider of the inocula, one would make an extrapolation, though this is preliminary, that we are seeing well in excess of a four log reduction in prion material.

This was done both in phosphate buffered saline, but also phosphate buffered saline that was supplemented with various concentrations of fetal cap serum or bovine serum. And interesting enough, in mock experiments that were not exogenously spiked with this inocula, looking at this process maintained the biological integrity of the bovine serum both in cell growth potential in terms of cell viability, in duplication rates, et cetera.

Why conceivably should this work? Well, as I mentioned before, there is a lipid component that is associated with PRP. The GPI-anchor is very, very important in terms of PRPSC misfolding. Interesting enough also, there are two cytosine residues which form a disulfide bridge or disulfide link which is highly susceptible to oxidation and maintains the secondary structure of PRPSC which is germane to its infectivity.

Also, you have a number of asparging residues. As you know, PRPSC is found in three forms, in unglycosylated or mono-glycosylated and di-glycosylated form. And these highly branched glycosyl radicals are highly susceptible to oxidation, and have been at least reported to be associated with chaperon proteins, in particular protein X, which has not been isolated yet, has been associated with infectivity.

Certainly this experimentation is preliminary. It will be continued, further experiments delineating the methodology of the mechanisms underlying this phenomena. Obviously, we will continue work with mouse brain homogenates, which is a little bit closer to the infectivity model. We will work with murine bioassays. We will do some biochemical analysis of bovine serum.

Interesting enough though, it appears that we will start preliminary work as well with plasma proteins, as well as fresh frozen plasma to evidence whether or not ozone can inactivate infectious prion proteins. This is the first technology that has been reported to actually inactivate -- not remove say through a filtration technology -- but actually inactivate prion proteins in a biological serum while maintaining its integrity.

Thank you.

DR. ALLEN: Thank you very much.

At this point we move to open committee discussion. It's 12:45 pm. I doubt that we can complete our discussion. We are supposed to wrap up by 1:00 pm and break for lunch. I think rather than starting the session and having the FDA perspective and questions for the committee and then breaking for lunch, because if we don't break at 1:00 pm, we aren't going to get lunch, I think we'll just break right now.

I would like to start again at 1:30 pm.

[Whereupon, the meeting was recessed for lunch at 12:45 pm, to reconvene at 1:30 pm.]

A F T E R N O O N S E S S I O N (1:35 pm)

DR. ALLEN: Before we get into the FDA position summary and questions, I would like to point out that Dr. Quirolo had found the new presentation of the Cochrane Collaboration. I'm going to ask him to give a two minute summary of that from what he has looked at. But Pearl has offered to provide it for committee members if you would like to have it.

Can I see a quick show of hands? Okay, we will get it reproduced for people.

Do you want to go ahead with your summary?

DR. QUIROLO: What I think I'll do is I'll just read their conclusions. And this was just published on the Internet, and it was published online on October 18, 2004. The authors' conclusions state:

"For patients with hypovolemia there is no evidence that albumin reduces mortality when compared with cheaper alternatives such as saline. There is no evidence that albumin reduces mortality in critically ill patients with burns and hypoalbuminemia."

"The possibility that there may be highly selected populations of critically ill patients in which albumin may be indicated remains open to question, however, in the view of the absence of evidence of a mortality benefit from albumin, and the increased cost of albumin compared to alternatives such as saline, it would seem reasonable that albumin should only be used within the context of well concealed and adequately powered randomized clinical trial."

DR. DI MICHELE: Can I just ask for a clarification? Did they say now -- did they reverse what they said about burns as well?

DR. QUIROLO: They actually didn't reverse it. But with the SAFE trial, when they redo their analysis, they say there is no difference between albumin and saline, with that caveat that I just read.

DR. DI MICHELE: But did they include burns though? Did I hear you say burns?

DR. QUIROLO: They used all the same studies they used in 1998, plus a few more. But the SAFE study was so huge, it just sort of overwhelmed the rest of these smaller studies.

DR. ALLEN: They apparently give equal weight to each patient enrolled in each one of the studies. And on that basis, the SAFE study just overwhelmed it.

Okay, why don't we move on to FDA perspective and questions. Dr. Landow, are you providing a summary?

Agenda Item: I. Safety of Albumin Revisited - D. FDA Perspective and Questions for the Committee

DR. LANDOW: Have data from the SAFE resolved the safety concerns that were raised in the meta-analysis by the Cochrane group for: (a) critically ill patients in general; and (b) subgroups of critically ill patients with burns, hypovolemia or hypoproteinemia?

Agenda Item: I. Safety of Albumin Revisited - E. Committee Discussion and Recommendations

DR. ALLEN: Okay, why don't we look at Question A first, have the concerns been resolved for critically patients in general? Comments and discussion.

DR. SCHREIBER: I guess I came away with a slightly different view. And that is that my impression from the presentations and a number of papers that I found that did additional meta-analysis that there is really not a lot of substance behind the Cochrane. So, we are basing everything on that particular review now, and we have heard repeatedly I believe, or I heard that there were some serious questions about the interpretation.

So, is it really that we are addressing the safety concerns of that? Because if those safety concerns are in fact not right, then we are coming through it from the back door.

DR. ALLEN: I think that's an interesting perspective, and certainly one to be considered. My understanding is that the FDA really has a couple of options open to them. If, depending on the committee vote, they could leave their current statement, which is still available on the Internet, on the Web site, they could leave that intact.

They could take it down without replacing it with anything. Or as the industry has asked, they could replace it with a statement that sort of goes the other direction if you will, or at least indicates that there aren't concerns. And we are not going to specifically instruct the FDA what to do. We are going to give them basic recommendations from the committee based on the information we have heard and what we have read, which then will lead to their decision.

DR. KLEIN: I would like to concur with Dr. Schreiber. I was one of the people who was rather outraged when the Cochrane study was published, both by the analysis and by the conclusions. I think FDA did what was prudent at the time, given the fact that even the editors came out with a very strong statement way beyond that I think reasonable people could conclude from the data that were presented.

I think now we have additional data. Even Cochrane has reversed much of what they said. But I still disagree with their conclusion, because I believe that they have again, once extended the data that are available, to say that albumin somehow is less satisfactory, even though that isn't their conclusion.

I think we are more or less back to where we were before, and that is the SAFE study has shown us that under the circumstances of that trial, there is essentially no difference. And so, I would think that we have been kind of blinded by the conclusions of the original Cochrane study. We should now take that off the table, and simply say there is no credible evidence that one is superior to the other.

DR. ALLEN: And it seems to me that one could go a step beyond that when you get to the B part of the question, and that is that there are suggestions that if one were to design and carry out adequate studies for subgroups of people, that we would be able to help physicians make much better judgments in the future.

The data simply are not available today to enable that, although there are suggestions that in fact there may be differences. And my feeling is although it's not asked for, I as a side comment would suggest to the FDA that to the extent -- and the NIH -- to the extent that appropriate studies should be considered in this area, that it probably would help clinical judgment in terms of these critically ill patients.

Dr. Lew.

DR. LEW: Actually, I was just going to make that comment. I think that what we have seen today is there are tantalizing data, but not definitive. There might be some subgroups that do better with albumin, and potentially there might be even subgroups that do worse with it. But in general as a group, you just can't say albumin is bad, not yet. And so, there are going to be other issues with the clinician and what they think the patient needs, and as well as costs to make those determinations at this time.

DR. KLEIN: Jim, I just wanted to add I concur with you that more studies are clearly needed. My concern, however, right now is that given the data that we have, physicians make decisions based on individual patients. They look at them, they evaluate them, and they treat them given the best available information.

I wouldn't like physicians who are treating patients in these various categories to start with the presumption that albumin could cause harm, because I don't think those data are there, and we have suggested that may be. So, I would like to again suggest that somehow that be taken off the table, and we start back where neither one appears to be superior, and you use your clinical judgment to determine what particular treatment you apply to an individual patient.

DR. HUDSON: Essentially, I agree with that. That the answer to 1A is yes, that the SAFE study indicates that in a heterogeneous group of critically ill patients, the two are equivalent in terms of efficacy or safety, as far as we can tell.

I think we need to be careful in terms of B, because I think one of the very positive things that the SAFE investigators and Dr. Finfer said was they have been very careful about even their a priori defined subgroup analyses. And I do think that the one group that probably concern should be raised about would be trauma with brain injury.

But having said that, point out that further studies about trauma and sepsis need to be done, and that otherwise there aren't enough data to say that one or the other is efficacious in any group of patients, with that one exception to be cautious about, even though that needs to be studied further as well.

DR. ALLEN: Good, thank you. Other discussion?

DR. KUEHNERT: This point at least tangentially has been made, but there is a risk on the downside to having these safety concerns or expressing these safety concerns where they don't necessarily exist, especially in terms of the patients with severe sepsis, where albumin could be beneficial. So, we also have to look at the risk of having this sort of a document out there. There is a risk to that also.

And I'm struggling a little bit with the wording on the questions, and hopefully maybe we can resolve that, because I think there was a specific issue here that maybe was not right to start with as far as the Cochrane analysis. And now in fact we have data that actually needs to be expressed and updated.

So, at least we are not in the awkward position of saying well, this analysis may not have been right, because it was not done in the right way. We can say, well, if you combine the data that now exists, you reach a different conclusion, regardless of whether that analysis was done properly or not. So, I think that makes our job a little easier.

The other thing I wanted to say is as far as the new Cochrane conclusion, I don't understand why they put cost considerations in there, where as far as I understand the meta-analysis didn't have anything to do with cost. So, that's a little confusing to me, especially since there are some countries where it is free apparently.

DR. ALLEN: Right. I think that's a very appropriate comment also.

DR. HUDSON: Just in terms of 1B, I think there we have to say one, the data from the SAFE study doesn't apply to patients with burns, because they weren't included. And we don't know enough about hypovolemia or hypoproteinemia to really say anything. Patients with those conditions would have been included in the overall study, but we don't know anything about those individual groups. And I think the previous data are just inadequate.

DR. ALLEN: Are we ready to vote on 1A? My interpretation, and rather than trying to rewrite, I'm satisfied with the question as a vote, with the understanding by the FDA that they need to take the vote of the committee in context of this whole discussion, and I think they will do that.

DR. DI MICHELE: I would agree with you for Question A, that I think the question as it's framed makes sense. I think for Question B, I'm not sure that it does. I think we might want to answer that question in a different way. But we can do that when we come to it, if you wish.

DR. ALLEN: Yes, let's get 1A taken care of, and then we can move to a more detailed discussion of 1B.

Dr. Epstein, are you satisfied with that approach?

DR. EPSTEIN: Sure.

DR. FREAS: Okay, I will go around the table and call off the names of the voting members. There are currently 12 voting members sitting at the table. Dr. DiMichele.

DR. DI MICHELE: The answer to 1A is yes.

DR. FREAS: That's correct. We are voting on 1A only. Dr. Kuehnert.

DR. KUEHNERT: Yes.

DR. FREAS: Dr. Klein.

DR. KLEIN: Yes.

DR. FREAS: Dr. Schreiber.

DR. SCHREIBER: Yes.

DR. FREAS: Dr. Lew.

DR. LEW: Yes.

DR. FREAS: Dr. Quirolo.

DR. QUIROLO: Yes.

DR. FREAS: Dr. Allen.

DR. ALLEN: Yes.

DR. FREAS: Ms. Baker.

MS. BAKER: Yes.

DR. FREAS: Dr. Hudson.

DR. HUDSON: Yes.

DR. FREAS: Dr. Manno.

DR. MANNO: Yes.

DR. FREAS: Dr. Doppelt.

DR. DOPPELT: Yes.

DR. FREAS: Dr. Whittaker.

DR. WHITTAKER: Yes.

DR. FREAS: As acting industry representative, Dr. Katz would you like to voice your opinion?

DR. KATZ: Had I been given the opportunity, I would have voted yes.

DR. FREAS: Dr. Leitman, would you like make a comment?

DR. LEITMAN: I would also have voted yes.

DR. FREAS: Thank you very much for your comment. That's a unanimous yes.

DR. ALLEN: All right, we'll move on discussion of 1B, which is have the data from the SAFE study resolved the safety concerns raised in the meta-analysis by the Cochrane group for subgroups of critically ill patients with burns, hypovolemia, or hypoproteinemia? Discussion.

DR. LEW: I agree with Dr. Hudson, that I think there is some interesting data that we -- like any other medication in this world, there are potential side effects, and sometimes the potential side effects to subgroups. And I think we have seen some data that looks worth looking at.

And I think the original statement made by the FDA made it clear that it's within the context of the clinician looking at the data, and coming up with their own conclusion. Maybe we can do something like that in here, although they specifically say subgroups, burns, hypovolemia, and hypoproteinemia, I would say in general that there may be subgroups that may do better with albumin, might not do better. But that is within the realm of people reading the literature and making their decisions.

DR. QUIROLO: As a pediatrician I have to mention that there is one subgroup that is not included, and that is infants and children. So, there are no studies to show efficacy here, although I do believe that in terms of the equivalents, I think they are okay. But there are no studies to show that they are okay for children and infants.

DR. LEW: But it was my understanding that in the Cochrane study they included some neonates, which I don't think is appropriate either, because they are totally different, the NICU.

DR. ALLEN: Other discussion? May we rephrase the question in the way that Dr. Lew suggested?

DR. FREAS: Would you please rephrase the question for us, Dr. Lew?

DR. LEW: I think just more general subgroups of critically ill patients, and leave out specifics.

DR. ALLEN: We would simplify it to say have data from the SAFE study et cetera, resolved the safety concerns that were raised in the meta-analysis by the Cochrane Group for subgroups of critically ill patients.

DR. LEW: No, that doesn't do it. I guess my concern is not necessarily in the context of this Cochrane Group study. I think enough has been said about it that we are unhappy with that study.

DR. ALLEN: Yes, in some ways what we need to do is get rid of the first part of it, and say are there sufficient data from the published studies in the literature to resolve safety questions for subgroups of critically ill patients?

DR. KATZ: I understand what is trying to be done here, but those concerns came from the Cochrane review primarily, so I think we are just playing with words a little bit here, I really do. I like it the way it is.

DR. LEITMAN: Would it be helpful to phrase the question such as are there subgroups of critical ill patients for which residual safety concerns exist? Because from what I heard this morning, there is one subgroup in my mind for which the safety concern is so real, that it merits mention, trauma with brain injury.

DR. ALLEN: Yes, and it's an interesting question of whether it's a safety risk or an efficacy risk, because -- well, it's difficult to really say, because what you are doing is you are comparing one substance versus another. What are the best protocols to use under what circumstances? And we don't have hard data on that.

Dr. Epstein, you were going to make a comment?

DR. EPSTEIN: Yes. The categorization of subgroups of burns, hypovolemia, and hypoproteinemia was within the original Cochrane analysis. That's why we have framed the question this way. We understand that the SAFE study did not address all those three subgroups. And what we are really getting at is whether there is enough evidence, either from review of the Cochrane report or from the SAFE study to negate the apparently stated concerns.

So, I think maybe the confusion here is arising from focusing it only on SAFE. In other words, if we said do the available resolve the safety concerns that were raised in the meta-analysis by the Cochrane Group, we would be better positioned. In other words, it may have been confounded by focusing the entire outcome of 1B on the SAFE study.

So, I would suggest that if the question is to be revised, and that's at the discretion of the committee, it would be do the available data, which would include re-analyses, other meta-analyses, published reports, and so forth, resolve the safety concerns that were raised in the meta-analysis by the Cochrane Group?

DR. ALLEN: And I could see a situation, depending on caveats, in which I could answer it either way. That's part of the problem.

DR. LEW: Just to try to address what Jay is saying. If all FDA wants to know is this committee's opinion about the Cochrane study, and then of course new data coming behind it, I would think most of us would agree that yes for A, as well as for B in specific reference to the Cochrane study.

I think my comments were kind of a broader -- and maybe that's not what you want. If that's not what you want, fine, we can be very narrow.

DR. ALLEN: That's why I said I could, depending on the caveats, answer it either way.

DR. LEW: I asked for instruction from Jay, if that's what you want, something narrow?

DR. EPSTEIN: Well, I think we do want a narrow question answered about the Cochrane study, because it was the basis for FDA alerting physicians to exercise caution. So, I think we are happy to have additional opinion expressed on where the field stands in general. But I think we do want a narrow answer whether we should now relevel the playing field with respect to the Cochrane report.

DR. QUIROLO: Well, I think what you should do then is say that this is the 1998 Cochrane study, because the 2004 Cochrane study says that albumin and normal saline are equivalent, reviewing all the previous studies, and adding in the SAFE study and a couple of others.

DR. KLEIN: It seems to me that if you wanted a narrow interpretation from me, I would say that the available data suggest that the only subgroup of concern is that of trauma with brain hemorrhage. I don't have enough data to talk about any of these other subgroups, but that would be the only one, if you wanted a narrow interpretation, that I would have any sufficient concern about now. And that is in fact based on the SAFE study.

DR. ALLEN: Although again, but it's not a group that is included in here, but if you've got somebody coming in with septic shock, as has been pointed out, you may want to preferentially use albumin. There is a lot of question about a lot of different subgroups I think. And it's not captured within the three that are included here, which were included because they were in the original, the 1998 Cochrane study. We are kind of going around a little bit.

DR. DOPPELT: Can I suggest that we go back to one of Dr. Lew's earlier statements? And that is, looking at various subgroups, there may be differences in their response to albumin or not. And that is something that needs to be further studied, but at this point you can't make really any conclusions unless you wanted to pick the brain injured as one subgroup. But other than that, you just don't have enough data to say one way or the other.

DR. KUEHNERT: I think there is a difference here I see in the question. We are trying to answer is there really an efficacy difference with these subgroups? But the question seems to be about the 1998 Cochrane meta-analysis, and specifically commenting on analysis. And I'm hearing narrow interpretations. So, I'm just wondering if we are trying to be too broad in trying to look at the data, when actually the question is just about the FDA Web site response to this meta-analysis.

DR. ALLEN: Yes, if I view it based on the clarification that Dr. Epstein gave, looking at this fairly narrowly, I would be comfortable voting have data from the SAFE study, or are available data, have they resolved the safety concerns that were raised in the 1998 meta-analysis by the Cochrane group for subgroups of critically ill patients with burns, hypovolemia, or hyperproteinemia? I can answer yes to that.

And then I would go on to add a summary statement for the FDA that there are still many questions in terms of the most appropriate use for colloid versus crystalline fluids for critically ill subcategories of patients.

DR. KUEHNERT: My only problem with that is the burns part, because they excluded them.

DR. KATZ: I won't be given the opportunity to vote, so let me just say that it's easy to say yes to A, because that's really the crux of the SAFE study. I would be very uncomfortable with burns, but I think we have a document out there from the FDA that served a reasonable purpose at the time. It is arguable, but it served a reasonable purpose to alert people to disturbing findings, but that's what was in the literature at the time.

And I believe that that document needed to be edited, and the discussion and the data that we have heard today needs in some way to be referenced in that document. But very clearly as an infectious disease doctor, if I'm going to use albumin as a primary resuscitation fluid in a septic patient, I'm not convinced by the data at this time.

I would want to see that done in an appropriately designed trial using state-of-the-art care in 2005. And whether that includes recombinant protein C or not, or whether it's glucocorticoid replacement or what it might is a very complicated issue. And I don't think that was reflected in the SAFE study.

So, it may be that that point estimate of benefit for sepsis in the SAFE study wouldn't hold up with the current approach to resuscitation from sepsis. It needs to be trialed. And I think that the FDA needs to change what is on the public record to reflect those uncertainties.

DR. HUDSON: I guess what I'm thinking in terms of B is that burns weren't included in the SAFE study, but if we are talking about the 1998 Cochrane analysis, that this data wasn't sufficiently strong enough to be able to say that albumin was clearly unsafe in those three categories.

DR. ALLEN: Yes, I think that's good.

DR. HUDSON: And then I think we just need to be very -- I agree entirely that sepsis needs to be further studied, as does trauma, but with the caveat and the statement that Dr. Lew has already mentioned, that people need to interpret the literature, but they also need to be aware of that one subset that needs further study as well.

DR. ALLEN: Dr. Epstein, are you satisfied with the committee's discussion, or do you want a vote on a reworded 1B?

DR. LANDOW: I think a vote is in order. But if I could just point out, and maybe people don't agree with this, but I thought hypovolemia was answered by the SAFE study.

DR. ALLEN: Burns were not.

DR. LANDOW: So, aren't there data to say that we have resolved the question of hypovolemia?

DR. HUDSON: Well, what we have done is resolved the question about a heterogeneous group of critically ill patients. And we don't know specifically, and I would ask Simon, I don't think we specifically know about hypovolemia, because it wasn't defined either well in the studies that the Cochrane analysis included. And it wasn't a specific subgroup defined in the SAFE study.

DR. FINFER: I guess the data that I presented in NUSA(?) last week is in the public domain. And that is that for patients with a low baseline serum albumin, defined as a baseline serum albumin level of less than 25 grams per liter, there is limited evidence for a treatment effect favoring albumin. So, on that basis, if you are asking my opinion, I would say that the SAFE study has answered the question about hypovolemia.

It has answered the question about hypoalbuminemia, although that data is not published. And I could not show that data, because the percentage points might change by 0.1 or 0.2, and it wouldn't be fair to the final publication, and it's disreputable to give you not the absolute right data.

We didn't include burns patients. My personal view is that that's based on three studies -- three poor, old studies. The SAFE study has shown that the data that was used to generate the concerns about hypovolemia certainly, and I would suggest also hypoalbuminemia has proved to be flawed. And then it's the committee's decision whether they want to extend that to say that three old, small studies should not cloud our view on burns.

I would still feel quite strongly -- obviously, you'll get our further analysis of the trauma with brain injuries group, hopefully at the end of this year. But as I said, in my practice I don't give those people albumin until I see that data, because there is no evidence of efficacy.

DR. SCHREIBER: I believe I heard Jay say that we didn't necessary have to focus only on the SAFE study. And maybe what we could do is add a couple of words that would say something to the effect of data from the SAFE study and other sources resolved the issue. And then I think I could vote yes for B.

DR. ALLEN: What I would suggest, do currently available data, including those from the SAFE study, resolve safety concerns raised in the 1998 meta-analysis by the Cochrane for subgroups of critically ill patients, including those with burns, hypovolemia, or hypoproteinemia?

DR. LEW: Number one, a correction. It should be hypoalbuminemia. And then the other is that you had mentioned earlier, and that will be fine if you don't do it, but if you had an addendum mentioning our general concerns about subgroups. I think it's also worth putting an addendum expressing our concerns about the Cochrane study of 1998 in the first place.

DR. ALLEN: I have no objection to doing that. What I would suggest is that we not make that part of the question, which I think if we do that, it gets just too complicated. But that we do that as a committee statement that goes along with our vote, if that would be satisfactory.

DR. HUDSON: I guess the only problem I have with that is burns, and that none of the new data deal with burns. I think we need to have a separate statement that says the data is very limited in relation to burns, and was not addressed in the SAFE study. But I agree entirely, the data that they presented initially was just so limited in both the size of the studies and the results that you can't say anything from those data definitively.

DR. ALLEN: Okay, let me suggest this wording. Do currently available data including those from the SAFE resolve safety concerned raised in the 1998 Cochrane Group meta-analysis for subgroups of critically ill patients with hypovolemia or hypoalbuminemia? And remove burns from it, because that wasn't in the SAFE study, and we really didn't discuss that in detail today.

We will add a committee statement at the end saying that the Cochrane analysis for burn patients was based only on three limited studies, which are now old. And that there are very intriguing data raised by currently available data, including the SAFE study, for other subgroups of patients that suggest further well designed, well controlled studies should be undertaken to resolve.

DR. DI MICHELE: I just want to express an additional concern. I wouldn't feel comfortable making a decision about the hypoalbuminemia, with all due respect, but based on data that I didn't see yet. It may leave some residual ambiguity. But then I can abstain.

DR. EPSTEIN: I think it would be helpful if we had three separate votes, burns, hypovolemia, hypoproteinemia, because there is no other way to really get a sense whether there is a consensus or lack of consensus of the committee. And we certainly appreciate all the expressed caveats. The discussion is extremely helpful. But I think it would help us further to go on the record with three votes.

DR. ALLEN: All right, Dr. Freas, are you ready to do three separate votes?

DR. LEW: I do want to add to the last comment, even in the presentation criticizing the Cochrane report it is pointed out if you look at all the studies on hypoalbuminemia, they are all skewed towards suggesting that albumin may not be ideal. And having been involved in meta-analysis before, it is always tantalizing when you have that many studies, and they all skew toward one side. And again, I think it was an excellent point to mention that we haven't seen the data yet. It would be nice to see the data.

DR. ALLEN: And if we don't want to vote either a clear or a clear no, we can always abstain. And that simply means that the data are not available. We aren't able to make an adequate on that. And that's not necessarily preferred, but that is a very acceptable solution that is available.

DR. KLEIN: Jim, I absolutely agree with you on the availability of the data, and we haven't seen it. But what is on the table now is that the use of albumin is somehow toxic. And I think we don't really have any data for that either. As you said, we have a number of studies that may fall in this area, but we have heard all of the reasons why these aren't really good for a meta-analysis. They are heterogeneous. They different kinds of albumin, et cetera, et cetera.

And I guess my only question is do you want to leave that open? Do you want to leave the thought that this is still not just an area that hasn't been resolved, but that albumin is potentially a bad thing to use here?

DR. ALLEN: And that's why I used the term advocacy earlier, because if you do nothing for these critically ill patients when they come in, trust me, the mortality rate is going to be higher than using albumin. I think your point is well taken.

Okay, are we ready to vote? We will run through and do three separate votes. The leading statement is do critically available data, including those from the SAFE study resolve safety concerns raised in a 1998 Cochrane Group meta-analysis for subgroups of and we will do separately burns, hypovolemia, hypoalbuminemia.

DR. FREAS: Voting solely on the subgroup burns at this time. Dr. DiMichele.

DR. DI MICHELE: No.

DR. FREAS: Dr. Kuehnert.

DR. KUEHNERT: No.

DR. FREAS: Dr. Klein.

DR. KLEIN: No.

DR. FREAS: Dr. Schreiber.

DR. SCHREIBER: No.

DR. FREAS: Dr. Lew.

DR. LEW: No.

DR. FREAS: Dr. Quirolo.

DR. QUIROLO: No.

DR. FREAS: Dr. Allen.

DR. ALLEN: No.

DR. FREAS: Ms. Baker.

MS. BAKER: No.

DR. FREAS: Dr. Hudson.

DR. HUDSON: I'm confused. I liked it the way you worded it before. The answer is no, but the data doesn't say that it isn't convincing that albumin is not safe.

DR. ALLEN: That is confusing. And that's where an abstention may be the best answer too.

DR. HUDSON: Abstain.

DR. FREAS: Dr. Manno.

DR. MANNO: No.

DR. FREAS: Dr. Doppelt.

DR. DOPPELT: No.

DR. FREAS: Dr. Whittaker.

DR. WHITTAKER: No.

DR. FREAS: Industry perspective?

DR. KATZ: I have a genetic predisposition against abstention, so I would vote no with Dr. Hudson's caveat that we don't have it either way.

DR. FREAS: Would our consultant like to comment?

DR. LEITMAN: I feel the same way. I feel I can't vote no, because I don't like the question. So, I'm not sure that I want to place any confidence in anything in the initial Cochrane study right now from 1998. So, I don't want to leave it as a potential negative, as albumin as a potential negative. But it wasn't addressed in the SAFE study. So, I will abstain -- I can't vote, but that would be my vote.

DR. FREAS: Okay, of the voting members we have 11 no votes and one abstention.

DR. LEW: I think it's unfair the change the question midway, because I agree with Dr. Hudson. And I don't know who else might have also agreed. We are voting on something that is silly, because I think we are all agreeing the Cochrane study did not show that albumin was bad for burn patients. And here we are being asked did the SAFE study show that it wasn't bad, when it didn't even address burn patients. It's a bad question.

DR. EPSTEIN: We hear the reservations of the committee that the Cochrane study didn't establish a risk in burns. But it is clarifying to us to comment on the scope of the interpretation of the SAFE study.

DR. FREAS: Are we going to vote now on the subgroup hypovolemia? Okay, we'll go around the table. Dr. DiMichele.

DR. DI MICHELE: Yes.

DR. FREAS: Dr. Kuehnert.

DR. KUEHNERT: Yes.

DR. FREAS: Dr. Klein.

DR. KLEIN: Yes.

DR. FREAS: Dr. Schreiber.

DR. SCHREIBER: Yes.

DR. FREAS: Dr. Lew.

DR. LEW: Yes.

DR. FREAS: Dr. Quirolo.

DR. QUIROLO: Yes.

DR. FREAS: Dr. Allen.

DR. ALLEN: Yes.

DR. FREAS: Ms. Baker.

MS. BAKER: Yes.

DR. FREAS: Dr. Hudson.

DR. HUDSON: Yes.

DR. FREAS: Dr. Manno.

DR. MANNO: Yes.

DR. FREAS: Dr. Doppelt.

DR. DOPPELT: Yes.

DR. FREAS: Dr. Whittaker.

DR. WHITTAKER: Yes.

DR. FREAS: Our industry opinion?

DR. KATZ: Yes.

DR. FREAS: And our consultant, would you like to comment?

DR. LEITMAN: Yes.

DR. FREAS: This is unanimous yes votes for the voting members.

We are polling hypoalbuminemia. Dr. DiMichele.

DR. DI MICHELE: No, based on the published data so far.

DR. FREAS: Dr. Kuehnert.

DR. KUEHNERT: Yes, I think I can accept unpublished data presented from a reputable source.

DR. FREAS: That was a yes vote?

DR. KUEHNERT: Yes.

DR. FREAS: Dr. Klein.

DR. KLEIN: I would have to vote no, with the reservation that I have no confidence in the initial interpretation of the 1998 Cochrane study on hypoalbuminemia.

DR. FREAS: Dr. Schreiber.

DR. SCHREIBER: I'm in the same state as Dr. Klein, but I guess I'll have to vote yes.

DR. FREAS: Dr. Lew.

DR. LEW: I think I'll abstain.

DR. FREAS: Dr. Quirolo.

DR. QUIROLO: Abstain.

DR. FREAS: Dr. Allen.

DR. ALLEN: I'm going to tip to the yes side, with significant reservations also.

DR. FREAS: Ms. Baker.

MS. BAKER: I'll have to abstain on this.

DR. FREAS: Dr. Hudson.

DR. HUDSON: Yes, like Allen.

DR. FREAS: Dr. Manno.

DR. MANNO: I vote yes.

DR. FREAS: Dr. Doppelt.

DR. DOPPELT: Yes.

DR. FREAS: Dr. Whittaker.

DR. WHITTAKER: No, like Dr. Klein.

DR. FREAS: Okay, I have 3 abstentions, Dr. Lew, Dr. Quirolo, Ms. Baker; 3 no votes, and 6 yes votes.

And our industry opinion?

DR. KATZ: I too think that the unpublished data is from a reasonably reputable source, and only wonder if the data holds for people who are upright, as opposed to people in Australia, and would vote yes.

DR. FREAS: And our consultant, would you like to make a comment?

DR. LEITMAN: I also am impressed with the data unpublished by the SAFE and feel much confidence in that than I do in the Cochrane study, so I would have voted yes.

DR. FREAS: I hope we're done voting.

DR. ALLEN: All right, I think the FDA has their results from the vote. I am not going to restate again the caveats that have been laid out, but I think the discussion was a rich discussion that laid out areas of research that need to be done.

I think it's fairly clear that one, that the treating physician in critical care areas needs to assess each individual patient, and the data are not currently available to allow physicians to make the judgments that they need to. And that obviously can be resolved with appropriate research over time.

Thank you all. We actually accomplished that all in 45 minutes, so we came in under our allotted hour. Everybody can take a deep breath. We're going to move on to our post-lunch agenda at the present time. We've got a couple more committee updates. The first is by Dr. Mark Weinstein on international agreements and harmonization.

Agenda Item: Committee Updates - Update on International Agreements - Mark Weinstein, PhD, OBRR, FDA

DR. WEINSTEIN: The purpose of this presentation is to give you an update on contacts with international regulatory bodies, and information-sharing agreements that deal with issues involving the FDA where the Office of Blood Research and Review at CBER has significant input.

Interaction with some of these entities is relatively new. We believe that these contacts give us new opportunities to harmonize regulations and standards, exchange information that can speed licensure of products, improve the quality and safety of blood products on a global basis, and facilitate compliance and surveillance activities.

Now, CBER is a World Health Organization, a Collaborating Center Biological Standardization. In that capacity we have multiple engagements and contacts with WHO. Our laboratories are very active in providing candidate materials for standards, and working on standards development for in vitro diagnostic devices, plasma proteins, NAT testing, blood typing sera, and on projects related to TSEs.

This work includes organizing international collaborative studies to standardize material, and to analyze data. We also have the responsibility of storing and shipping some of these standards. Partners in these activities include: the National of Biological Standards and Control in the UK; the Paul Erlich(?) Institute in Germany, and Sanquin(?) in the Netherlands.

We work with WHO's Expert Committee on Biological Standardization as temporary advisors to establish candidate materials as international reference standards and reference panels. Now, ECBS has recently endorsed the formation of a task group of experienced regulatory authorities and technical experts in the blood field.

The idea of this new group is to work on scientific assessment of current threats and technologies that impact public health, come up with opportunities for cooperative action, and opportunities for regulatory harmonization. This group will also serve as a source of expertise to support WHO in assisting regional regulatory networks. FDA has been through CBER or through OBRR, involved in discussions on the formation of this group.

CBER also supports the global collaboration for blood safety. In fact, Dr. Epstein is the chair of this group. WHO is a participant of the GCBS and also provides its secretariat. The GCBS focuses on global blood safety and availability, and is composed of internationally recognized organizations, institutions, associations, agencies, and experts from developing and developed countries. It serves as a forum for international collaboration on issues of blood safety and availability through dialogue, non-binding recommendations, and cooperative work.

We have a number of opportunities to harmonize our regulations and standards with other regulatory bodies. And I put the word "harmonization" in quotes, because it has different means depending on the context. Now, in the context of the International Conference on Harmonization, we participate with regulatory authorities in Europe and Japan, and experts in the pharmaceutical industry in these three regions to discuss scientific and technical aspects of product registrations.

The decisions of this group are binding, and will be adopted by all three regions. I should mention that PPTA and the blood industry are not part of this process. This includes primarily the biotech industry.

Efforts to harmonize our standards are taking place in less formal situations and non-binding contexts. We have recently become an observer at the Group of Experts 6B of the EDQM European Pharmacopoeia. This group makes decisions about required assay procedures and potency reference standards that are used in Europe, and our participation as an observer gives us the opportunity to understand the rationale for their decisions and voice our opinions. We not bound by 6B decisions, but the discussions can help us as we formulate our own policies.

Also, the Office of Blood is very active in developing the recommendations that are in the annually updated guide to the preparation, use, and quality assurance of blood components that is issued by the Council of Europe.

We have many collaborative contacts with the National Institute of Biological Standards and Control in the UK. NIBSC plays a primary role in developing many reference materials, including organizing international collaborative studies for determining potency, analyzing data, preparing ampules for distribution, and shipping materials. It is our hope that this collaboration will be able to continue, even as the NIBSC is undergoing some reorganization within the health department in the UK.

We also collaborate with the scientific and standardization subcommittees of the International Society of Thrombosis and Hemostasis. The subcommittees work on evaluating potential candidates for potency reference materials, and the recommendations are submitted to WHO's ECBS committee for review. The subcommittee also makes recommendations about clinical trial design. They develop patient registries, and help in surveillance activities in areas such as inhibitor developments in patients receiving clotting factors.

Finally, I would like to point out the numerous information sharing agreements that we have with other regulatory bodies. These include: the European Commission, the European Medicines Agency of the European Union; the Health Products and Food Branch of Health Canada; SwissMedic in Switzerland; the Therapeutics Goods Administration in Australia; the Pharmaceutical and Food Safety Bureau in and Pharmaceutical and Medical Device Agency in Japan; and the Health Science Authority in the Republic of Singapore.

Now, this information sharing allows us to exchange non-public information about regulated products as part of cooperative law enforcement and cooperative regulatory activities. I should mention in particular one very positive example of information sharing that we recently received, and this involves the EMEA. They have invited us to comment on guidance documents that they are developing for clinical trial designs for Factor VIII and Factor IX. This represents another opportunity that we have to potentially harmonize our regulatory processes and improve product availability.

Thank you.

DR. ALLEN: Thank you. Comments or questions for Dr. Weinstein?

DR. KATZ: I'll be very brief. The efforts of harmonization that have occurred to date really largely exclude the blood community in the United States. And we have an enormous interest in harmonization, particularly with Europe, because so much of the recovered plasma that we supply to fractionators goes there.

And those blood centers that supply recovered plasma in particular to the European fractionators are in the process now of coming under European inspection protocols. And while some people might find this statement unusual coming from me, we find the FDA to be far more reasonable in their inspection criteria than we do the Europeans -- far more reasonable, enormously more reasonable.

And one of the deterrents to moving a lot of this harmonization discussion into the area of plasma, plasma derivatives and blood is that there is really no funding for that part of the effort from our tiny little minimum margin community.

DR. DI MICHELE: Although Dr. Katz's comment is certainly justified, certainly in terms of blood derivatives, I think this is really a step in the right direction. And I think with respect to a lot of the challenges that we face with new product development, et cetera, and especially in the rare bleeding disorders field, I think this is a very important move. And especially where we are dealing with rare events in rare diseases and such, and products for even rare disease I think this is very much a step in the right direction.

But I would second what Dr. Katz said, because I think that certainly the issue before us today does indeed involve harmonization with respect to plasma, plasma products, et cetera, and the classification thereof. So, some extension of that harmonization would certainly be even more welcome.

DR. ALLEN: Other comments or questions?

I agree, certainly this is much needed. And I know how frustrating it can be to participate in some of these discussions. They get rather ponderous, and trying to meet needs across borders is not easy, but I think it's wonderful that it's moving forward.

All right, our second update is sharing information with the public. The speaker is Ms. Kathleen Swisher from the Office of the director.

Agenda Item: Sharing Information with the Public - Kathleen Swisher, RN, JD, OD, FDA

MS. SWISHER: Good afternoon. I was asked to speak about sharing information with the public. And in particular, I am going to be talking about guidances and regulations.

First of all, I just listed some of the tools that are available to FDA in order for us to share information with the public. First and foremost is The Federal Register. In The Federal Register we issue notices of availability for guidance documents, and also draft proposed rules and final rules. And I'm sure most of you are familiar with these.

We also, in addition to documents that are published in The Federal Register, we share information with the public by holding workshops and meetings. And the amount of information that we can share during these workshops and meetings depends upon where we are in the phase from concept to actual document development.

These are some other tools that we utilize, and also some of the provisions. We are trying to post documents on the Internet, trying to use the Internet more often now to disseminate information. And this is still of course developing, but I think we are moving forward in this area.

We also have started opening dockets for comments to the dockets. I know we have done this one time as the result of a workshop where we didn't have enough time for adequate opportunity of comments during the workshop. So, we opened a document so that further comments could come in.

Our ability to share information with the public has been codified under Title 21 under the Code of Federal Regulations 10.80. Here it describes where during the concept phase or the idea phase we can share openly our ideas and our recommendations or what our thinking is. I don't know that I should use that term -- what our considerations are I should say, concerning a certain issue.

And this is where we like to communicate a lot with the public during workshops or meetings, or even with experts to gather information and data and recommendations from the public.

Once we decide to go forward with a document, either a draft guidance document, or a proposed rule, at this point in time we can no longer share the details with the public. We can only share our general ideas. And again, this is actually codified under 10.80.

Let me go back a little bit to sharing the details of the document. An example of this is about a year ago we published a final rule, the bar code rule, that some of you might be familiar with. And during the proposed rule stage when we were drafting the proposed rule, we could share with the public that we were drafting a proposed rule that would create a uniform bar code system to prevent medical errors. So, this is the general concept of the proposed rule.

What we could not disclose during the drafting of the proposed rule was the fact that we were considering putting the national drug code number into the bar code, and requiring that that would be part of the bar code. that's a detail that we could not disclose, because it would be disclosing a deliberative process that were undergoing as an agency. And this would be a pre-decisional document. So, we would not want to share those details with the public at that point.

Another reason that we can't share details with the public is -- where 10.80 enters into sharing with the public is that we must create a level playing field. And that is described here.

I just wanted to also give you some of the access information as far as Web sites and that sort of thing. Also, we publish two agendas. One is an annual guidance agenda. This is published in The Federal Register. And this describes our guidance documents that we are developing, and also those that we're thinking of revising.

We also publish a semiannual regulatory agenda, and this describes any regulations that we were working on, and that are in development, or again for revision. We do accept comments on these, and I know that I have received comments on a proposed rule that I have been working on that was published in the regulatory agenda. And we do consider these comments when actually drafting our regulation.

These are the citations that I referenced: 10.115 are the Good Guidance Practices, and 10.80 is the dissemination of information. And the reason we connect the guidance documents with 10.80 is because a level one guidance document, according to the GDP's, must public with a notice of availability, and they are in The Federal Register. So, that's where we make the regulatory link, I guess you could say, to 10.80 in the dissemination of information for draft guidance documents.

These are our contacts. This is my staff on the top, regulations and policy staff. And we also have the Office of Communication, Training and Manufacturers Assistance. They also help us with our Web site and are responsible for the Web site. So, please if you have ever any questions, please contact us.

Thank you.

DR. ALLEN: Thank you very much. Questions or comments? Do you find that there are difficulties sometimes in resolving issues under the Freedom of Information Act with the need during the draft guidance development for relative secrecy?

MS. SWISHER: We do. Actually, we're very responsive under FOIA requests as much as possible. And my office doesn't deal with those. I have dealt with those when I worked in the Center for Devices, but I can't answer from any personal experience.

Thank you.

DR. ALLEN: Thank you.

We will move now into discussion under open committee discussion, Topic II. Review of Standards for Plasma Products for Transfusion. This will be introduced by Dr. Weinstein.

Agenda Item: II. Review of Standards for Plasma Products for Transfusion - A. Introduction and Review of the Literature - Mark Weinstein, Phd

DR. WEINSTEIN: This session will be focused on review of the standards used for making products for transfusion. We are asking for the committee's advice on the extent to which available scientific data may support changes to further standardize processing of plasma products for transfusion, and whether there is a need for additional scientific studies that would help resolve current areas of uncertainty.

So, what is the background of this request? The current standards, as set forth in the CFR and the AABB Circular of Information poorly define and characterize plasma products for transfusion. As we shall discuss this afternoon, there is scientific uncertainty as to the extent to which manufacturing conditions of plasma products affects the final properties of these products.

We are in the process of evaluating data from multiple sources to consider the possible development of minimal standards for these products that would further insure their clinically relevant safety, purity and potency.

At the workshop on plasma standards that was held in August 2004, a number of parameters were identified that could possibly affect the quality of products for transfusion. These include the time and temperature of separation of plasma from cells; the anticoagulant used; the freezing and storage, thawing and post-thawing conditions.

This afternoon we will review some of the key literature on plasma processing. I will point out some areas where improvements might be made -- not must be made. Following my presentation, Dr. Irma Szymanski will talk about her clinical experience with these products.

So, we will talk about some of the safety and quality issues associated with plasma product preparation. We will talk about the types of products produced, and their labeled indications, potential areas of improvement, and finally a major focus of this discussion will be on freezing rate as it affects fresh frozen plasma and products derived from FFP.

So, let's discuss what's the problem. For one, many of the standards in the CFR were derived more than 40 years ago, not that there is necessarily anything bad with deriving things 40 years ago. They could still be true. But in some cases, these standards are the result of expediency and practicality reflecting the technology of 40 years ago, rather than having any scientific rationale.

A number of side effects and hazards associated with the use of plasma for transfusion are outlined in the AABB Circular of Information. These include: immediate and delayed immunological complications; transmission of infectious diseases or infectious agents; circulatory overload; hypothermia; or metabolic complications.

Given these side effects and hazards, it is prudent to make these products in a way that will minimize exposure to them and their potential risks. And one way to do this is to try to maximize the potency for their intended use.

So, I'll describe the current situation with respect to products and their usage. Here are some product definitions. FFP can be made from plasma that is kept in contact with cells for up to 8 hours at room temperature, or 1 to 6 degrees, and from whole blood or by plasma pheresis. It is to be placed at -18 degrees centigrade or lower within 8 hours, or within the time specified in the directions for use for the blood collecting, processing and storage system, and stored at -18 degrees or lower.

The AABB Circular further specifies that if the blood is collected in ACD rather than CPD, it should be placed at -18 degrees or lower within 6 hours.

Now, as we shall see later, these allowable variations in conditions can lead to products with very different properties, but all of these products would be labeled FFP. These variables again include time and temperature of plasma contact with cells, the anticoagulant used, residual cellular content, and particularly the rate of freezing.

The rate of freezing can be affected by the volume of plasma to be frozen, the shape of the container holding the plasma, and the freezing device such as cold room, freezer, or a blast freezer.

Now, FFP can used as the source material for a number of different products including: cryoprecipitate; plasma, cryoprecipitate reduced; and thawed plasma. The idea again is we are using this material as a source for the production of other materials.

Cryoprecipitate is prepared by thawing FFP between 1 and 6 degrees centigrade. The AABB Circular says that each unit of cryoprecipitate should contain at least 80 units of Factor VIII and 150 milligrams of fibrinogen. The CFR says that AHF should be made by a procedure that produces an average of no less than 80 units of antihemophilic factor, and does specify fibrinogen content.

Plasma cryoprecipitate reduced is prepared from FFP that is thawed and centrifuged with cryoprecipitate removed by centrifugation. And does this requirement for centrifugation restrict potentially better ways of making cryosupernatan?

Thawed plasma is derived from FFP thawed at 30 to 37 degrees, and maintain at 1 to 6 degree centigrade for up to five days. I point out that this is not a licensed product.

However, there is a licensed product called Liquid Plasma. It is plasma that is to be separated from whole blood no later than 5 days after the expiration date of whole blood, which means it could be sitting with whole blood for 21 or 35 days, depending on the anticoagulant. It is to be stored at 1 to 6 degrees centigrade.

Now, considering the possibilities of cellular breakdown products entering the plasma, as well as proteolysis and degradation of the plasma proteins, one has to wonder about the safety and quality of this product. And the question is, should this continue to be a licensed product?

Plasma frozen within 24 hours after phlebotomy is plasma that must be separated and placed at -18 degrees or lower within 24 hours of whole blood collection. This product is frequently made from blood collected on mobile units. Again, the allowable time and temperature of plasma contact with cells can vary greatly.

Regarding the labeled uses of FFP, the labeled uses are the broadest of all plasma for transfusion products. It includes: replacement for multiple coagulation factors in bleeding patients; reverse of warfarin therapy; transfusion for TTP patients; replacement for factors for which there are no concentrates or recombinants; and for the management of rare plasma protein deficiencies such as C-1-esterase inhibitor and Factor V.

Now, as I have mentioned, there is also a source material for other plasma products. These identified uses of FFP suggest that it should be made in a way to retain the activity of labile proteins such as Factor V, Factor VIII, and von Willebrand Factor, non-labeled proteins, and activities that are poorly understood like those that reverse the symptoms of TTP.

Now, even though cryoprecipitate is not recommended as a first choice of treatment for hemophilia or von Willebrand's disease, it is nevertheless available for that purpose in an emergency situation. Considering the risks, it will be prudent to make it in a way that would reduce the number of necessary infusions.

It is also used in the practice of medicine as a sort of fibrinogen. And again, it would be advantageous to have a high concentration of protein per unit of cryoprecipitate as possible to avoid unnecessary use of the product. In the case of cryopoor plasma it is thought to be advantageous to have reduced level of von Willebrand factor. As we shall see later, under certain methods of preparation, von Willebrand factor content in cryosupernatan can be a function of the rate of freezing.

Twenty-four hour plasma, thawed plasma and Liquid plasma have all been considered in labeling to be equivalent to FFP, except as a source of Factor V and Factor VIII. And as we go along, we may in fact look at those conditions, look at the properties and material, and find that they perhaps are not equivalent.

Now, what are some of the factors again that affect the quality of transfusible products? It's well known that blood drawing technique and mixing of anticoagulant can have a significant impact on the quality of plasma. If done properly, there is the potential of activation of the coagulation system and other proteins. Although this is obviously important, these are not the major topics of this discussion, that has focused more on the physical and chemical parameters of product preparation.

The type of anticoagulant used can affect the quality of plasma that is stored above freezing for a given period of time. Residual cells that remain in plasma after centrifugation have the potential of causing immunological reactions, and the release of proteolytic and other cellular products even though the products should be ABO compatible. Potentially, some of the hazards and risks already mentioned are caused by cellular residues in plasma transfusion products.

Cellular content would be affected by centrifugational methods employed, which can vary with regard to the number of spins, and the centrifugational force.

The time and temperature from phlebotomy to freezing for FFP in the US is a maximum of 8 hours, unless otherwise specified in the directions of the collection device. The Council of Europe recommends that plasma from whole blood be frozen within 18 hours, with the optimal time being within 6 hours. And again, I say it's worth considering the scientific rationale for these differences, and whether the longer time is justified by data, and should we consider it for our own products?

One important point to bear in mind is that conditions that may be optimal for preservation of plasma protein activities, may not be optimal for cellular component preservation. Some, but not all investigators recommend chilling blood rapidly to 6 degrees or less to preserve red cells. Platelets are better preserved at 22 degree.

These two different pre-freeze conditions can produce different levels of Factor VIII and von Willebrand Factor. And any consideration that you may have about revising standards has to take into account the total picture, and not the issue of preserving plasma protein quality.

The temperature of plasma that is in contact with cells can strongly influence the concentration of Factor VIII and von Willebrand Factor. The so-called cold activation of the coagulation and kinin systems, and I will discuss each of these a later more fully.

Finally, the quality of transfusible plasma products can be strongly affected by the rate that plasma is frozen, and the thaw and post-thaw conditions under which cryoprecipitate and cryosupernatant are produced, and I will return to these topics later.

But let's look at some experimental data. This is an experiment that demonstrates the loss of Factor VIII and von Willebrand Factor that can occur while plasma is in contact with cells at low temperature. Whole blood collected in citrate tubes in this experiment were centrifuged immediately and plasma frozen at -80 degrees. The sample serves as the control for the experiment.

Other aliquots of whole blood were held at 22 degree or 4 degrees for 3.5 hours before separation of the plasma from the cells. The plasma was tested for throbinin, Factor V, VII, VIII, IX, X, XI, XII, and von Willebrand Factor antigen and von Willebrand Factor collagen binding. And only Factor VIII and von Willebrand Factor were affected by whole blood kept at 4 degrees.

In this experiment there was an average loss of 25 percent of Factor VIII, and about 50 percent of the von Willebrand Factor in these samples if kept at the lower temperature. So, whether the same degree of loss would occur in plasma collected in CPD or ACD is not clear from these experiments, but this experiment does suggest that some loss will occur. And again, this material is the starting material, the source material for the production of cryoprecipitate and other materials. So, one has to wonder at this stage, we are losing significant amounts of Factor VIII and von Willebrand Factor.

Another point where improvements could be made is illustrated by these experiments. This graph shows the effect of anticoagulant on plasma stored at 6 degrees for up to 35 days. And as you can see if I bring up peptide A, it is released over this period of time in both the CPD and the ACD plasma, with anti-thrombant-3 remains relatively flat.

The study noted that these storage conditions can result in activation of the coagulation, fibrinolytic and calacrin systems. You might not that Liquid Plasma is a licensed product, and can be separated for up to 5 days after the expiration of whole blood. And whole blood can be stored up to 35 days, depending on the anticoagulant. And so, one has to consider is this an area where perhaps we should reconsider the licensure of this particular product?

Now, as I noted in the introduction, FDA held a workshop on plasma standards last year to review the scientific data, regulatory requirements and current industry practices regarding the freezing, storage and shipping of plasma. Among the objectives of this workshop were to help develop standards for recovered plasma, and to examine the potential of harmonizing our regulations with those in Europe.

This chart shows you what comparison of the US standards for FFP and the Council of Europe recommendations for plasma for transfusion. I draw your attention to the respective conditions of freezing, storage and expiration date, shipping temperature and allowable deviations. We require plasma to be placed in a freezer at -18 degrees or lower with storage at that temperature, and then have an expiration time of 1 year, a shipping temperature not to exceed -18 degrees. And we don't allow any deviations from those storage and shipping conditions.

The Council of Europe recommendations include freezing to -30 degrees within an hour, and storage conditions of -18 to -25 degrees for up to 3 months. If the material is stored below 25 degrees it has a storage period of 24 months.

The question is, is it really necessary to freeze within an hour? Or to have an expiration of date of one year? And these are obviously conditions that can affect our own regulations. Should we consider changing some of these conditions as they are supported by scientific data?

In considering establishing new minimal standards, it is important to consider the scientific basis for our current standards, and advantages and disadvantages of rapid freezing. We also need to be mindful of the current use and need for plasma for transfusion products, as well as the practical implications of modifying current standards.

So, we have already talked about these particular conditions where plasma derived from whole blood can be affected by these elements before plasma is frozen. And now we will talk more specifically about the effects of the rate of freezing on the quality of plasma products.

This is from a study by Carlebjork in 1986. It shows the change in temperature over time as 700 mLs of plasma is frozen to -70 degrees. And note that a significant amount of time is spent in the phase transition portion between 0 and -2 degree centigrade. I want you to pay particular attention to this time period, because I will be referring to it later on in some of our further discussions.

The second panel shows the amount of time that it takes for the 750 and 1,500 mLs of plasma to freeze under varying conditions. This is in a -25 degrees freeze box, a -80 degree freezer, and at -40 degrees in a circulating coolant situation.

And as you can see, it takes almost 10 hours for a 750 mL bag of plasma that is placed in the -25 degree freezer to freeze. But it takes a little more than an hour when this material is put at a -40 degree circulating bath. In the case of 1,500 mLs it takes up to almost 24 hours for the material to freeze.

The fact is that I have not seen any studies examining what happens at -18 degrees. And the point is that these freezing times can be changed dramatically depending primarily on the ability to remove heat, and the temperature of the container. So, again, this idea of freezing at -18 degrees versus to a given temperature, which signifies that it's the rate of freezing that is important.

So, what happens to the product under these conditions, or at least the Factor VIII under these conditions? This graph demonstrates a rather dramatic loss of Factor VIII activity in freeze frozen plasma as a function of the time it took for the phase change to occur. And as you can see, almost 60 percent of the Factor VIII activity can be lost if a phase change takes 5 hours. And again, current regulations do not specify the rate, and a considerable amount of Factor VIII activity can be lost because of this slower freezing process.

This is another example, another experiment demonstrating the effect of slow and rapid freezing on plasma. And in this case we used 200 mL aliquots of material. The sample was prepared within 2 hours of blood draw. And aliquots were frozen at -40 degrees within 40 minutes. This was classified as fast freezing. Or slow frozen in a cabinet freezer at -40 or -20 degrees.

The time to phase change was not reported in these experiments, but the samples were tested after 14 days of storage. And you can see the losses in Factor VIII activity that occurred under both conditions. Here there is a slow freeze at -20, and a slow at -40. Under both conditions it wasn't the final temperature that resulted in the loss of material, but it was the rate of freezing that had a major effect compared to the conditions of rapid freezing, again both at -40 degrees.

So, rapid freeze under these conditions, there was about an 8 percent loss of Factor VIII activity, while the slow freeze at -20 degrees, there was a 20 percent loss of Factor VIII activity, and 14 percent at -40 degrees centigrade. These are under the slow freeze conditions.

There was also an 8 percent loss of Factor V, and again, these conditions not only affect Factor VIII, we are talking about other labile proteins like Factor V. Again, the freezing rate is more important than the temperature of the freezing device.

In this study the release of beta-TG from residual platelets was also examined. In the rapid freeze there was a 9 percent release of beta-TG, which was not statistically significant, but a 14 percent increase in the slowly frozen material. This observation indicates that rapid freezing might not only preserve more Factor VIII content, but also disruption of residual cells that may be plasma. So, it's not just the Factor VIII that we are talking about. We are talking about the fact that freezing rate can break up cells if it is done slowly, in contrast to the more rapid freezing.

Should we be concerned about residual cellular content of plasma transfusion products? And are they responsible for some of the transfusion reactions that we have observed?

Other parameters that were measured in this experiment include: Factor VII, von Willebrand's Factor, soluble thrombin, anti-thrombin complex, and C-1-esterase inhibitor. There were no changes in the measured quality or activity of these proteins between fast and slow freezing.

We will next turn to the effect on cryoprecipitate and cryopoor plasma of freezing and thawing. This experiment uses 200 mLs of plasma that were frozen in a -40 degree freezer, and at -70 degrees in a liquid nitrogen ethanol bath. I don't know how many blood centers might have a -70 degree liquid nitrogen ethanol bath, but these are pretty severe conditions.

In this case the phase change at -40 degrees took about 1.5 hours, while it took about 8 minutes for the plasma to be frozen at -70 degrees. Under the conditions of this study there was not a statistical difference in total Factor VIII clotting activity, determined as the sum found in cryoprecipitate and cryosupernatant between the slow and fast freeze conditions.

This suggests that under the conditions of this study, Factor VIII in FFP was minimally affected by rapid or slow freezing. However, the freezing conditions did significantly affect the recovery of Factor VIII, von Willebrand's Factor, and fibrinogen in cryoprecipitate and cryosupernatant. And you can see the higher levels of the Factor VIII and the von Willebrand Factor in the cryoprecipitate compared to the cryosupernatant.

So, what do we derive from this experiment? Under these rather extreme freezing conditions, 25 percent more Factor VIII, 24 percent more von Willebrand Factor, and 13 percent more fibrinogen were in cryoprecipitate made by the fast freeze method compared to the slow freeze method. Potentially less cryoprecipitate and cryopoor plasma prepared by the fast freeze method would be needed to treat the labeled indications for these products.

Another set of data, however, emphasizes the importance of specifying the conditions under which FFP, cryoprecipitate, and cryosupernatant are made. In this study, 200 mL samples of plasma were placed in a fluorocarbon circulating bath and reached -27 degrees in 20 minutes. Samples were also placed in a -30 degree cold room, and it took about 120 minutes to reach -27 degrees. The phase change time was about 45 minutes for the slow freeze, and about 5 minutes for the fast freeze.

Upon thawing, taking this fresh frozen plasma and thawing it, at 37 degrees there was 4 percent less Factor VIII activity in the slow frozen method compared to the fast frozen. But you can see that over time this difference disappeared, and that these are the hours post-thawing; within 24 hours they lost about 40 percent of the Factor VIII activity as the plasma just sat at the 22 degrees.

So, what happened to the cryoprecipitate that as made from the slow and fast frozen plasma? The FFP was stored at -8 to -12 degrees in a cold room over 18 hours and fully thawed at 4 degrees in a water bath. The cryoprecipitate was separated by centrifugation, rather than the thaw siphon method that was used in the previous study.

The important point here is that there was no significant difference in the Factor VIII, von Willebrand Factor, or fibrinogen content between these two conditions. And so, what I'm emphasizing here is that the method used to make cryo was different in these two different situations, and that this can be a major determining factor in the amount of Factor VIII and other products that you get from these products here. It's very method-dependent.

So, again, what are we to make of these results? We need to carefully define what we mean by fast and slow freezing conditions, specifying the rate of freezing, and particularly the time taken for phase transition are important parameters that should be measured. It appears that increasing the freezing rate of plasma up to a point decreases the loss of Factor VIII content in FFP beyond which further increases in the rate has little effect.

Importantly, these conditions may not be far from the current industry procedures. And I think that this is one of the major points here. What is the current industry standard? How is plasma actually being made? How far are the conditions of -30 degrees within -- well, I should mention this last point here.

The amount of Factor VIII, von Willebrand Factor and fibrinogen in cryoprecipitate is highly dependent on the thaw method, as well as the freezing rate. Under defined conditions there is little advantage in freezing to -30 degrees in 1 hour compared to 2 hours.

And again, we may have a situation that currently exists within industry that would permit these conditions to occur. These in fact may be close to what the standards are, the present actual standards of manufacturing fresh frozen plasma. And that is something that we should be looking into and find out about.

The final topic that I will discuss will be the issue of storage time and temperature. Currently, FFP has an expiration date of 1 year, and what is the justification for having plasma stored for only 1 year? In 2000, Kotitschke published the results of an experiment in which FFP was prepared from plasmapheresis plasma. This is pool 1. And whole blood donations obtained at different centers, and this is pools 2 and 3.

The plasma was placed in 200 mL bags. And pool 1 was frozen to -30 degrees within 35 minutes, pools 2 and 3 at -50 degrees within 30 minutes. The samples were stored at the given temperatures and times that we have here. And assays were performed by 13 different laboratories. Tests included determinations of Factor VIII activity, Factor IX, V, fibrinogen, thrombin, anti-thrombin, pro-thrombin, F1-2 fragments, anti-thrombin and C-1-esterase inhibitor.

The authors concluded there was no significant changes in any of the measured proteins over 24 months between plasma stored at -20 degree versus -40 degrees. And this observation is in agreement with findings made by Farrugia, who found no change in Factor VIII or fibrinogen in cryoprecipitate made from plasma stored for 6 months at -20 degrees versus -40 degrees.

Kerner(?) also found no change in small aliquots of plasma shock frozen at -25 degrees, and stored at -20 and -40 degrees for 9 months. He tested for Factor V, VIII, IX, thrombin, AT-3-plasminogen, plasma, pre-calcarine, and calcarine, and again, found no change.

So, are conditions of freezing, of storing fresh frozen plasma for one year justified, or is there room where we might think that perhaps there is room where we might change some of these standards?

The only problem, however, is that there is some contradictory information that we have to take into consideration. In this study small aliquots of plasma were frozen at -74 degrees or -24 degrees and tested again for multiple factors including thrombin, 5, 7, 8, 9, 11, 12, protein CS, anti-thrombin, plasminogen, von Willebrand Factor, and D-diren(?). This shows just the result of the Factor VIII decay.

And here we see samples that were frozen. The circles here represent material that was frozen at -24 degrees, and stored at -24 degrees, versus that that was frozen at -74 and kept at -74. In this particular experiment there was a considerable loss of Factor VIII activity in these small aliquots over a period of time.

And that's one of the issues that we have to consider there. We have conflicting information. What are we to take as being supportive of particular evaluation of conditions and potential changes in our standards?

The final issue I would like to mention here is the study that shows the effect of changes in temperature on stored fresh frozen plasma, affecting fibrinogen and von Willebrand Factor content and Factor VIII content in cryoprecipitate. FFP stored at -40 degrees was warmed to -5 degrees for 1 day, and then back again to -40. And then again, the temperature was raised, and then dropped again to -40.

Here we see the effect on Factor VIII content. It turns out that under these oscillating conditions, Factor VIII activity really didn't change very much, but there was a considerable increase in the fibrinogen content of cryoprecipitate made again under the thaw siphon conditions.

So, an important question, to what degree is this result a function of the thaw siphon method, as opposed to the centrifugational method of preparing cryo from cryosupernate, and whether one gets the same result by conditioning the plasma before completing thawing it 4 degrees. But this shows you there is a method in fact of increasing the fibrinogen content in your cryoprecipitate. Considering the cryoprecipitate is actually used a great deal as a source of fibrinogen, this is important information, and again, should be considered in our standards development.

So, what can we conclude? Most, but not all studies suggest that rapidly frozen FFP can be stored at -20 degrees for more than 1 year without significant changes in the activity of certain proteins, including Factor VIII. Current standards specify 1 year storage of FFP.

Temperature fluctuations in frozen FFP can affect cryoprecipitate composition, and current regulations for FFP do not permit deviations above -18 degrees. The question is should one or more of these standards be reconsidered?

So, a few more caveats when considering evaluating these scientific studies that might support updating our standards. A comparison of study results was complicated because of differences in measured parameters and assay techniques. And you have to be very careful when you look at these various studies, are we comparing apples and oranges?

Laboratory studies may not reflect practically achievable conditions, and there we might consider the -70 degrees in the liquid nitrogen ethanol bath conditions here. But they might not be necessary either.

Conditions optimal for production of one product may not be optimal for others. And again, we have to consider the total picture, the production of cells as well as plasma proteins.

So, requests to the committee, these are issues that you will be discussing. Please discuss the extent to which the available literature on plasma processing may support changes to improve the clinically relevant safety, purity, potency or consistency of various plasma products for use in transfusion such as the time for plasma separation, time from collection or separation to freezing, freezing rate, and target temperatures, storage temperature, allowable temperature excursions, and cellular content.

Secondly, what addition scientific studies are needed?

Thirdly, what recommendations do we have for the next steps forward?

Thank you.

DR. ALLEN: Clarification questions for Dr. Weinstein?

DR. KATZ: Mark, on several occasions you used the word "quality" with regards to the product at the end. And I just want to make the point that we don't know what the quality is, because the kind of trials that are required to establish optimal quality have never been done. And it's not so much quality that is being measured, but individual parameters that may or may not affect the clinical outcomes from infusion of any one of the many products, or the amount required to reach a clinical endpoint in a variety of those sorts of considerations.

DR. DI MICHELE: Mark, at first in trying to reconcile all these data, it seems that even in looking at any one single parameter, things are very confusing. We have heard this data a few times. Maybe internally within the FDA this data has sort of been looked at again. And I know since that workshop you have actually had a chance to look at some of the other literature, which you have incorporated into this talk.

Is there anything that we can sort of begin to hang our hats on so to speak, with respect to any physical, chemical, properties of freezing and storage, et cetera, that you noticed, or that sort of come to light through this review of the literature that provide a good starting point for future studies, let's say? Maybe internally --

DR. WEINSTEIN: Well, I think there are a lot of interesting questions that one could ask. For example, this business about the cold activation of plasma, the remarkable loss of Factor VIII and von Willebrand Factor that can occur in plasma as it is in contact with cells before it is frozen. That to me, is a very interesting observation, and should be looked at further. But it does raise the question of whether one should, for the purposes of making a defined product here, try to minimize that time of contact.

Another element here was this cellular content of plasma, the residual cellular content of plasma. And the idea that under fast freezing conditions, apparently there was less cellular lysis compared to the slow freezing conditions. These might have an effect on what happens to the patient who receives material that is frozen slowly versus that which is rapidly frozen.

Another area, to me, Liquid Plasma, it's amazing stuff here. I'm surprised that this material is still a licensed product. It seems that over the allowable time period, that you do get activation of the coagulation system, fibrinolytic system, there is a whole series of things that could happen within the 35 days. And that to me doesn't appear that this is a very good product to be allowed to be used up to a 35 day period.

Now, clearly the Factor VIII issue is very interesting. If you consider maybe not having to use five bags of cryo, if you could use 4 bags of it, you might be reducing the level of exposure of a patient to a product that -- unnecessary exposure to a product here. So, clearly the rapid freezing element here can preserve more Factor VIII.

And it may also have an effect on things that aren't measured. And I think that's another point here. This is an issue that was raised very strongly at the plasma workshop, that Factor VIII is a surrogate for things that we don't know about. And things that for example, are perhaps involved in the use of product for TTP. We think we know a little bit more about what is necessary there, but not entirely actually.

This is also a source of materials for those replacement of rare factors, of rare disease Factor V. We saw it in some of the data here that Factor V activity was also lost between the fast and slow freezing conditions.

So, I think that there are a lot of elements that bear looking into further. However, I think it was also interesting that the -30 degrees, this idea of you can freeze it rapidly up to a certain point, and then it doesn't really matter.

If you can use the -30 degree -- if that experiment holds true, we believe what the literature says here, the difference between a fast freeze down to -30 degrees within 1 hour doesn't make much of a difference compared to if you freeze within 2 hours. You will get the same result. And again, this is at -30 degrees, not an unreachable temperature given the current technology.

DR. LEW: I just wanted to comment, when you look at your challenges to evaluating data, and then your considerations, just looking at the two studies you mention about Factor VIII and changes in it after storing fresh frozen at different degrees, they were both different types of studies. Since I really don't have the N to each one, I can kind of count the points for one of the studies, but others I can't.

I guess I'm still concerned that you say most, but not all studies suggest that it can be stored. This is kind of a leap, isn't it?

DR. WEINSTEIN: I wish the data was flat out clear. That's really the point that I'm raising here. That it would be very nice if all the papers agree that yes, you can store fresh frozen plasma at -20 degrees for 2 years. It wouldn't matter if you stored it at -20 or -40, it's all fine.

I have to say that the studies that Kotitschke did were with the same kind of volume of material. That is, 200 mLs of plasma were being stored. In contrast to these small 5 mL aliquots that were being stored in this other study. So, it is apples and orange? I don't know, I'm just raising the issue here that one might be concerned about making blanket statements.

DR. LEW: I guess to add to that though, just looking at the numbers you see a huge difference between the -20 in terms of both low levels and high levels, and then a very tight level at the -40. In just my own experience with biological products, lower degrees you actually preserve activity much longer when it's been in the lab for years and years. And now, significance in terms of clinical --

DR. WEINSTEIN: Are we talking about the Kotitschke? Is that what you are looking at?

DR. LEW: Yes.

DR. WEINSTEIN: Okay, well, the important point to now there is that these are different centers, or different aliquots, different pools of material that were being analyzed. They weren't the same. So, you are seeing a group, pool 1, pool 2, and pool 3 are really different materials that were being tested in each of those situations.

DR. LEW: No, I understand different pools.

DR. WEINSTEIN: But you see the tightness of the result?

DR. LEW: Yes. There are some things about it that you have to question.

DR. WEINSTEIN: You do have to look at the whole -- I agree, one has to be very careful in assessing what the significance of this study is. However, it does seem to be the most complete one that we have using the appropriate volumes of material and the storage conditions that are practical.

MS. BAKER: Thank you for that summary. Can you comment on the extent to which some of the international groups, the COE or some of the other bodies that you so nicely described your collaborations with in your previous presentation, are these addressing these?

DR. WEINSTEIN: Every meeting of the Expert Committee 6B has a discussion of freezing of plasma. And I can tell you that they go through the same agony that we do in trying to assess what the proper conditions should be. There are political considerations that come. One country doesn't wish to do something that another country does, and there are compromises that are made, and all the rest of that. But the same sort of considerations and worries and thoughts are going on there too.

MS. BAKER: Are they at the same level of discussion as we are here, where they are just beginning to raise these raises?

DR. WEINSTEIN: They have been raised there for actually a much longer time. And I think that they have gone about this business of looking at the science. And that is where they came up with 2/-30 degrees in 1 hour. That was their decision. I think 2/-30 degrees in 2 hours will get you the same result probably. But they looked at the science that would support that thinking.

The -18 degrees, why are we saying -18 degrees? Well, apocryphal comments here that that was what the temperature of the Sears freezers could go down to 40 years ago, and that was fine, and let's freeze it at that temperature. It's the idea that this was the technical capability 40 years ago, and we can do better now. And we are doing better now.

And we could probably validate what we are doing now, because many of the current situation is that people are using freezers that can freeze rapidly. It would be interesting to see the data from those kinds of freezers that are being used now. And I would be interested for manufacturers to see what they could come up for data to submit to us.

DR. MANNO: I was interested in your concern about cellular fragments that might be transfused based on freezing techniques in thawed fresh frozen plasma. Could you elaborate on that, please?

DR. WEINSTEIN: I can only elaborate on what my reading has. But I think that Dr. Szymanski might have some information about that, or the cellular content as being a concern in these fresh frozen plasma products. She will be discussing that.

DR. ALLEN: Okay, perhaps we should move on then to the second presentation. I'm sure we'll be asking additional questions later.

Our next presentation is by Dr. Irma Szymanski, Professor of Pathology Emerita, University of Massachusetts Medical School. She will be discussing clinical use of plasma for transfusion.

Agenda Item: II. Review of Standards for Plasma Products for Transfusion - B. Presentation (Clinical Use of Plasma) - Irma O. Szymanski, MD, Professor Emerita of Pathology, University of Massachusetts

DR. SZYMANSKI: Good afternoon, everybody. I hope you have enough stamina to listen to this again. Dr. Weinstein gave a detailed and interestingly controversial presentation about plasma. And I'm going to go over some of the same stuff.

My presentation is from the point of blood banks and transfusion services. And we are sort of concerned about any changes that might happen to plasma. That's very, very important to us, because we have to do a lot of changes then. However, although you don't like the word "quality," but I think the quality of the plasma components is really an integral part of the whole plasma transfusion therapy.

Now, the following products will be discussed, and from the outset I want to say that thawed plasma and Liquid Plasma are two products I have never used in transfusion. So, they will get very little attention. And I'm sure if I would try to give it to the clinical staff, you would refuse to get them.

Now, I'm going to present to you the current way of preparing components in blood bank. I'm going to start here with the whole blood. And there is a CFR regulation that says that if you prepare only red cells and FFP, you have to keep the whole blood at 4 degrees. And depending on the length of hold, you get different plasma products. FFP, if the hold is 8 hours, FP24. If the hold is 24 hours, and then this crazy Liquid Plasma that you can separate within 5 days of the expiration date.

If you also prepare platelets from whole blood, everything has to happen at 22 degrees for the sake of platelets. But then in the end product you also get FFP. Is this FFP different from the one prepared from 4 degree blood?

Interestingly, there is a sort of secret document in FDA. And that doesn't appear in the CFR. And that deletes(?) preparation with two components also from whole blood that is kept at 22 degrees. And I know that very well, because I was cited for doing just that by one FDA inspector.

This shows basically the schematic outdating of these plasma products. And since it has been discussed thoroughly, I'm not going to concentrate too much on it, except stating that most of the products can be products at 4 degrees for 24 hours except cryo. And in the UK they can be kept only for 4 hours before transfusion.

So, what factor are important in recovering all the plasma proteins and factors in plasma? Dr. Weinstein gave a very thorough presentation. I'm giving you a simple one. And I want to bring up one factor that has not been really considered and that is the rate of cooling from the body temperature to the temperature of the hold. I wonder if that's important? I don't have any answers to that question.

Now, when cryo prepared, it is said it depends how you prepare it. And I have here some data from different centers, and you can see that the best data comes from -- oh, I'm sorry, I went ahead of myself. So, this is basically showing the anticoagulants that we use to collect whole blood. And these are designed to preserve red cell viability. However, most of the plasma is taken out of the red cells before storage, so it remains with plasma, I wonder if it is optimal for plasma, particularly since plasma proteins do not metabolize what is a sugar therefore.

Now, this is the data of Dr. Farrugia again showing you lose Factor VIII if the whole blood is kept at 4 degrees rather than at 22. There is of course loss also at 23 degrees.

And this shows another kind of data from Dr. Hughes which shows a step-wise decrease of Factor VIII in both temperatures, but more at 4 degrees. And this is because Factor VIII is cryoprotein, so it precipitates. And you can get it back by warming the whole blood to 23 degrees before separation.

These are data from the British Journal of Hematology 2004, and these data show that after thaw there is a certain decay, a delayed biocoagulation factors, which you would of course guess. And there is more decay in Factor VIII than Factor V. Actually, the half-time of Factor V is 10-14 hours supposedly in vitro.

Now, the cryoprecipitate which Dr. Weinstein explained how it is prepared, is a mixture of cold insoluble proteins. And it depends on the technical variables how much of each you get. And CFR states you have to have 80 units of Factor VIII in each bag, that's all.

This is comparison of different cryos. And you can see that the areas in Pennsylvania has the best results for Factor VIII, for von Willebrand Factor, and fibrinogen. They are using flash freezing of the plasma and cryo, and subsequent storage at -20 degrees. Data for Factor XIII and fibronectin were not very accurate or were available.

The cryoprecipitate is transfused to congenital hypo- or dysfribrinogenemia, and also for congenital Factor XIII deficiency, although in Europe there is a product that is purified for that factor.

It is also used as a source material for manufacturing, and for Fibrin Glue, which is a biological tissue adhesive used by surgeons. And it is prepared by combining cryo as a source of fibrinogen with thrombin solution, with calcium and usually antifibrinolytic agents. Interestingly, there is a commercial preparation only available by Baxter, Tisseel VH.

Fibrin Glue is used more frequently in Europe than in the USA. One of its problems is that the patient might develop antibodies to Factor V and thrombin if bovine thrombin is used as one of its constituents.

Now, when you transfuse plasma, the ABO group should be identical between the recipient and the donor, or at least compatible. Is it said that Rh type does not matter. However, I have observed several cases where Rh negative persons who had previously developed anti-D actually get autoimmune hemolytic anemia for a period of time. It doesn't last terribly long. And if they get Rh positive FFP, and this is an example of a secondary immunization by the D antigen which apparently is relatively intact in the fragments of red cells that occur in FFP.

Then on the other hand I have not seen any data where primary immunization to these cell fragments would occur. And that antibodies for red cell antigens or (?) antigens would develop.

It was already mentioned that blood problems would affect blood transfusion reactions, which are classified if they are acute or on the basis of patients' symptoms as allergic, anaphylactic, febrile non-hemolytic, or TRALI. And often we don't have an exact cause for allergic or febrile reactions. So, it was suggested that maybe some changes in plasma proteins during storage would be responsible.

We don't know, but one thing I would like to tell you is that when I have been studying stem cell collections with this mixture of different kind of white cells, they metabolize very heavily at 22 degrees, not at 4 degrees. And they consume oxygen and they release CO2. Now, I wonder if at whole temperatures, particularly at 22 degrees, these cells don't get enough oxygen, and they might break down and release some of their contents, which might be responsible for febrile reactions.

Now, the question is are plasma products, particularly FFP, transfused correctly? In the data obtained in the UK and the US state that this is not so. According to these data about 30 percent of FFP is given outside the guidelines. That's interesting.

So, what are those guidelines? And for the purpose of this talk they are classified into two categories, A and B. In Category A the person is having one single coagulation factor or plasma protein deficiency, and FFP transfusions are given if there is no purified product available. In Category B the patient has developed multiple coagulation factor deficiencies due to a disease or its treatment.

The first one, Category A is congenital Factor V deficiency. FFP infusions are effective, but if the patient has inhibitors to Factor V, it is more complicated. And then platelet transfusions, because they contain Factor V are recommended.

The second deficiency is Factor XI deficiency. And this factor prevent clot lysis, and therefore if major surgery is conducted at the site where there is high fibrinolytic activity, antifibrinolytic agent should be given as well.

The third is TTP, thrombotic thrombocytopenic purpura, which is a very severe disease, and results in death in 90 percent of the cases if not treated in a timely fashion. But currently, it has a quick response to plasma exchange and plasma transfusions with 80-90 percent recovery and survival. It is thought to be caused by a deficient of the von Willebrand Factor metalloprotease, ADAMTS13.

Now, an important concern here is what is the best plasma product when you replace the plasma taken out? And it has been stated that cryosupernatant would be the best, because it has the least amount of von Willebrand Factor multimers. However, the recent data in 2004 showed -- actually, it's 2001, showed that FFP and cryosupernatant are equally effective in terms of patient survival.

The last single factor deficiency is C-1-esterase inhibitor deficiency which can cause angioedema. FFP is okay to prevent angioedema during certain procedures or treated. There is a purified concentrate Cetor which is not yet available here. Hopefully, it will be soon.

Then we go to the multiple coagulation factor deficiencies. The first one to be discussed is DIC, disseminated intravascular coagulation, which is caused by several triggers to the vascular wall. And it starts a hemostatic system problem, and coagulation, microvascular coagulation, thrombosis and bleeding, and the platelets are soon depleted.

Now, treatment is to treat the primary cause, but if bleeding is present then all these plasma components are good. But there is no need to treat if there is no bleeding.

The second one in this category is liver disease in which the liver is unable to synthesize certain clotting factors. And bleeding can occur during surgery, with a biopsy, or variceal rupture. And then transfusion of FFP, a question of FP24, because of the low concentration of Factor V, and cryoprecipitate is indicated.

Prothrombin complex concentrates have been advocated, particularly in Italy, but the ones available in this country can cause DIC, and many hematologists do not want to use them.

Now, clinicians don't want to do liver biopsy if there is a slight elevation of PT, and this supposedly is not dangerous. You can do the biopsy with a slight elevation of PT.

The third Category B consideration is massive transfusions. And in this situation there is a dilution of coagulation factors, and surgeons usually like to give FFP on schedule. And particular practice is not recommended.

Fibrinogen deficiency is the first deficiency to occur when both 150 percent of the plasma volume is gone. And if bleeding occurs, we can give all these factors. But follow the response by the coagulation tests.

Open heart surgery nowadays is developed so much at least people do not need any transfusions at all. However, in the post-operative period there might be excessive bleeding from the chest tube, and that could be due to surgery problems, and should be differentiated from coagulation deficiency by the blood tests.

I insert this, because I think it's interesting. We have observed that if there is a fibronectin deficiency occurring the open heart surgery, this line is a fibronectin concentration which is significantly lower than that of the control proteins C3 and IgG. I haven't seen many other situations when you actually would need fibronectin.

The last consideration is over anticoagulation by warafin. Warafin works by preventing vitamin K-facilitated carboxylation of these factors: II, VII, IX, and X. As well as decreasing anticoagulation proteins C and S.

According our hospital guidelines, FFP transfusion should be given only if INR is 10 or more, and/or bleeding is occurring, or a patient requires urgent surgery. Many people have actually recommended giving Novel 7, which is a recombinant activated Factor VIII to decrease bleeding, but it's not very well understood. The recent data in the New England Journal of Medicine showed that (?) can decrease bleeding in hemolagic stroke.

Summary. The purpose of plasma transfusions is to correct deficiencies in plasma proteins and coagulation factors which cause serious, even life-threatening symptoms. Now, if there is a need to change the methods of plasma component preparations, it is important that they are defined on the basis of current knowledge and accurate data. And one should compare the protein content in the proposed methods to those that we are having now.

In addition, there is the problem of overuse of plasma. Some clinical studies are probably needed to define when FFP transfusions are necessary, and when they are not necessary in order to agreement between clinical and laboratory physicians.

And to guide physicians in the FFP transfusion therapy. They need to know a little bit more of what these products contain, and hopefully be able to use point of care coagulation tests to know what is happening, because everything goes very fast.

So, thank you so much for your attention. And I hope you understood my Boston accent.

DR. ALLEN: Thank you very much, Dr. Szymanski. Clarification questions? It's four o'clock. We have got a break, an open public hearing with three speakers scheduled, and then committee discussion and recommendations for the FDA to complete this afternoon. What I would suggest is that we take questions for clarification of Dr. Szymanski, and then take our break, move into the open hearing, and then come back and invite Dr. Weinstein and Dr. Szymanski both to answer the questions that we then might want to pursue in terms of the investigation.

So, clarification questions for Dr. Szymanski at this point?

DR. DI MICHELE: Thank you for your presentation. I have two questions. Could you again clarify -- I know you said it, but just for purposes of clarification, when you went through the preservation of Factor VIII and von Willebrand Factor components. And you mentioned the American Red Cross, and levels that were highest in some of their preparations. And you mentioned their method of preparation. Could you go over that one more time?

DR. SZYMANSKI: Well, they use so-called flash freezing. That is -50 degrees freezers that they put the plasma, and it takes about 2 hours for FFP to freeze totally, and less than 1 hour for the cryo. I don't know if they prepared it from the whole blood or kept that at 4 degrees. I believe this would have been the case.

DR. DI MICHELE: And that is once you actually you separate it, that's what you're talking about? Once the plasma is separated, then it's flash frozen. But you don't know how long between collection and separation?

DR. SZYMANSKI: Eight hours is the maximum.

DR. DI MICHELE: Okay, just the usual. And then you said it was stored at -22? Is that it?

DR. SZYMANSKI: Minus 20 or something, -18 or -20. There are some secrets here. They don't want to tell me everything.

DR. DI MICHELE: Thank you. And my second question is based on the current indications as you have reviewed them for the use of fresh frozen plasma, the indication that involves replacement of the most labile of all factors is Factor V, is that correct?

DR. SZYMANSKI: Factor VIII is most labile. Factor V is the second, that is not that labile.

DR. DI MICHELE: No, I'm just wondering, because we don't use fresh frozen plasma for replacement of Factor VIII.

DR. SZYMANSKI: I know. I'm just saying that usually Factor VIII can go down also in other conditions when you have other factors going down. So, it's not just the hemophilia A treatment.

DR. DI MICHELE: So, your recommendation then or based on your presentation, Factor VIII, the addressing issues of lability of Factor VIII is still justified given the conditions under which we currently use FFP?

DR. SZYMANSKI: Yes, I think so. But on the other hand, I'm not terrible strong about that recommendation, because usually you can give cryoprecipitate in addition to the plasma that you give, so you probably do get enough Factor VIII in these situations.

DR. KATZ: In that same slide with the four blood centers, what is the N that was used in determining those activities? What's the basis of those? I presume they are means or medians.

DR. SZYMANSKI: They don't really say that. It's sort of published data what they say it is, but they don't say standard deviation or number of cases studied. Neither did ARC.

DR. MANNO: I'm still interested in your case of the Rh negative recipient of plasma with Rh positive fragments. Could you explain what happened with that patient, please? And then in investigating that, did you find anything more on cases like that?

DR. SZYMANSKI: Well, we have several cases. The first case actually was somebody who went to open heart surgery and had several antibodies, D and a couple of others. And we were very careful giving negative units of red cells for those antigens. And then the patient suddenly comes back. He has severe hemolytic anemia, severe.

And then I get all the lawyers coming to my office to try to figure out what wrong we have done. I was trying to think very hard, what happens. And then I realized that this patient has gotten Rh positive plasma. And I thought, oh, this is interesting.

And then we had a couple of other cases in a similar situation, that they first developed or had anti-D, then they needed plasma infusions. And nobody cared about Rh, and then they got this porsidic(?) wounds and hemolysis. And then they treated and they did very well. But this sort of brought to my attention -- is that enough? Do I need to continue.

Before that I had noticed that even if you once get immunized to anti-D, or even some other antigen, so, first of all you have not antibody after transfusion. Then you get the anti-D or some other antibody. And then later on you get positive groom's(?) test. And there are no cells that are positive for these antibodies. And those then were oral antibodies actually, because they reacted with oral cells when they were elevated(?) from them.

So, it's basically a similar thing. And I published these in the Transfusion as actually an abstract. It was a talk. Hopefully, I want to write it up, have enough energy to write it up.

DR. KLEIN: I have to say, there is a big literature, as I'm sure you know, on autoantibodies following red cell transfusion, especially in people who already have allos, and may be forming additional allos. They make autos. We have published on that. There is a large literature.

I've searched the literature for any evidence of primary sensitization even in women who might have been sensibilized as Molison(?) used to say, and have not found a single case where Rh positive fresh frozen plasma has immunized someone primarily. So, I think this is interesting, but I would be very surprised --

DR. SZYMANSKI: I was very surprised too.

DR. KLEIN: If it's an intact Rh antibody D that is causing this. In fact, you have yet to convince me. I think it's interesting.

DR. SZYMANSKI: When I write the paper, I will send it to you.

DR. ALLEN: Other clarification questions?

Thank you very much, Dr. Szymanski.

Let's go ahead and take a break.

[Brief recess.]

Agenda Item: Open Public Hearing

DR. ALLEN: Welcome back to the open public hearing. We've got, as I said, three speakers. I need to read the open public hearing announcement for general matters meeting statement again.

Both the Food and Drug Administration, FDA and the public believe in a transparent process for information gathering and decision-making. To insure such transparency at the open public hearing session of the advisory committee meeting, FDA believes that it is important to understand the context of an individual's presentation. For this reason, FDA encourages you, the open public hearing speaker, at the beginning of your written or oral statement, to advise the committee of any financial relationship that you may have with any company or any group that is likely to be impacted by the topic of this meeting.

For example, the financial information may include the company's or group's payment of your travel, lodging or other expenses in connection with your attendance at the meeting. Likewise, FDA encourages you at the beginning of your statement to advise the committee if you do not have any such financial relationships. If you choose not to address this issue of financial relationships at the beginning of your statement, it will not preclude you from speaking.

Our first speaker for this afternoon is Dr. Michael Fitzpatrick, America's Blood Centers, Chief Policy and Chief Operating Officer. He will be speaking on rapid freezing of plasma for transfusion.

DR. FITZPATRICK: Good afternoon, thank you.

I am employed by America's Blood Centers, and so receive all my compensation from our members, and that's my disclosure.

There are just few slides here, and we'll go through them fairly quickly. We are speaking on the impact of a change in freezing temperature on the sites that do the freezing. We fully support the statement that is going to be provided by Kay Gregory later on AABB on clinical impact.

Hopefully, you know who we are by now, so we won't go over that in detail. We do represent all the non-Red Cross, not-for-profit community blood centers in the country. And we ship over 250,000 liters of plasma from volunteer donors for manufacture into plasma therapeutics annually.

We surveyed our blood centers and hospitals prior to the workshop that was held by FDA on freezing of plasma to determine what is currently being done, and what would the impact of a change be. This is just an idea of what we did in that survey: identify the products that they produce and store; describe how they freeze it.

Hospitals were asked what products do you hold on hand, at what storage temperatures? And they were asked what would you need to modify current temperature settings or purchase new freezers?

Fifty-two blood centers responded, about 68 percent of our membership. And that covers 75 percent of the supply that we collect, 5.7 million units; 168 hospitals that are served by our members responded of approximately 3,300 that get blood from ABC facilities.

To look at what blood centers manufacture, you can see here fresh frozen plasma; plasma frozen; cryoprecipitate; recovered plasma; and other, the preponderance of it being fresh frozen plasma.

How long do they hold it before they place it in the freezer? Here about 13 percent for 12 hours; just less than 40 percent up to 24 hours; and then a few at 72 hours and some other times.

Storage practices. Storing at -18 degrees centigrade, which is the current practice, about 52 percent of the blood centers, 67 percent of the hospitals as reported by blood centers. We make that differentiation, because we had response from the blood center telling us what their hospitals did, and then we had direct responses from the hospitals themselves, and that's here.

Twenty-two percent of the hospitals that directly responded -- let me rephrase that. Of the hospitals, 22 percent directly responded, and they store at -18 degrees, 68 percent at -25 degrees, 7 percent at -80 degrees. And then 19 percent that the blood center responded at -25 degree, none at -80 degrees. So, the storage temperatures are -20, -18, and -25. And we chose those temperatures, because those are what were being discussed at the workshop and proposed by FDA.

What's the impact of a change? Changing from -18 to -20 degrees may in some instances still mean we have to reset your alarm settings, because you want the alarm to go off in time for you to recover or rescue the product before it gets above -18. So, your alarm might now be set at -20, -21, so you hold the temperature at -18.

If you are going to change the storage temperature to -20, you are going to try and move the alarm down from -21 to probably -25. Some of those alarms have to be reset by the manufacturer. A technician has to come in and reset the alarm in the freezer and adjust. Sixty-nine percent of our centers would have to have that done by the manufacturer, 41 percent of the total hospitals responding.

Twenty-seven percent could do it themselves, 48 percent could do it themselves, and as you can see, some weren't sure. And sometimes you're not sure when you look at the freezer, whether you can actually reset the alarm and validate it yourself, or whether you have to bring in the manufacturer.

The cost would range from $500-1,000 per freezer, and with the number of freezers on hand, the average for blood centers would be about $16,700. For hospitals, $200-5,000 per hospital, average about $1,087.

If you have to purchase new freezers, which many would have to do in order to freezer at -50 degrees or do blast freezing to increase the rate of freezing as Dr. Weinstein was talking about we'll that 40 percent of the centers would have to purchase freezers, another 40 percent would not have to, and 20 percent weren't sure. For the hospitals 19 percent said yes, they would have to, 67 percent said no, and 14 percent weren't sure.

The cost here, $2,000-280,000, with an average of $75,000 for the blood centers. For a hospital, $3,000-30,000, with an average of $8,000.

So, in summary, about half store plasma at -18, half store in freezers capable of -25. Most freeze between -20 and -24 right now for recovered plasma, not fresh frozen plasma; 69 percent said that if they required storage at -25 degrees or below, they would have to have a contractor change the alarms settings; 27 percent no resetting necessary. Cost estimates, $500-100,000, average of $16,000, and then an average of $75,000 for new equipment.

For the hospitals, 68 percent store in freezers capable of -25; 41 percent would need to change the alarm settings, 48 percent would not have to. Cost estimates there were an average of $1,000 per hospital. Sixty-seven percent would need new freezers. And of 41 respondents, the cost of new equipment was the averaging of about $8,000 or $9,000 per hospital.

So, there is an impact financially to the centers. There is an impact work-wise to validate and reset temperature settings on freezers, not a huge impact, but there is an impact. And so, from this survey we just want to make you aware of the impact of a change, that it's not as simple as turning a dial on a freezer. There are other things involved.

And to change the freezing rate significantly, which would mean requiring that -50 blast freezer, that there is a bigger impact there with the purchase of new equipment.

Thank you.

DR. ALLEN: Thank you very much. Any clarification questions for Dr. Fitzpatrick?

All right, our second speaker is Ms. Kay Gregory, Director of Regulatory Affairs, AABB on the issue of plasma products for transfusion.

MS. GREGORY: Thank you. My salary is paid by the AABB, so that's my statement.

And Mike already indicated that ABC supported the statement I'm about to give you. I should let you know that the American Red Cross also supports the statement as I'm going to give it. And in the interest of time, and because I believe most of you know the AABB, I'll skip who we are, and get right to it.

In preparation for this meeting the AABB consulted its Clinical Transfusion Medicine Committee or CTMC for those of you who have a written copy, to determine how various plasma components are being used in clinical practice and if changes in preparation of these components would improve clinical outcomes.

Plasma is most commonly transfused into patients with multiple coagulation defects. The most common uses are for patients undergoing surgery and experiencing surgical blood loss, trauma patients, patients with impaired clotting resulting from liver disease, patients on excessive coumadin therapy, and treatment of patients with thrombotic thrombocytopenic purpura.

For treatment of these conditions, the Factor VIII levels in the transfused plasma are not critical to the therapeutic effect of the component. Specifically, Factor VIII levels in patients experiencing general surgical bleeding are usually above normal, since Factor VIII is an acute stress response protein.

In patients with liver disease, Factor VIII levels are generally normal. Factor VIII plays no role in coumadin reversal. The CTMC noted that Factor VIII is critical only in the case of treatment of Factor VIII deficiency. In that case, plasma components would be used to treat Factor VIII deficiency only in the rare instance that factor concentrate is not available.

Fresh frozen plasma, plasma frozen within 24 hours of collection, and thawed plasma are used interchangeably in many facilities, especially at large tertiary care hospitals with busy Level I trauma centers. One of the physician members of the CTMC also reported using these components interchangeably in a busy therapeutic apheresis center.

There are, however, some specific treatment indications where FFP and FP24 is used. An example is the treatment of neonates with coagulopathies. In other words, thawed plasma is not generally used for neonate. In summary, when used appropriately, FFP, FP24 and thawed plasma are all widely used and can produce good clinical outcomes.

AABB believes there is no current problem with the efficacy of plasma for transfusion. In the absence of new data, therefore, there is no clinically compelling reason to require changes in the preparation and storage of plasma components. In fact, much of the literature reviewed at this meeting attempts to judge the quality of plasma for transfusion through the measurement of Factor VIII levels in components. As noted above, Factor VIII is generally not clinically relevant to the therapeutic benefit of plasma components.

In addition, AABB cautions that Factor VIII is not an appropriate surrogate for the measurement of levels of other more stable coagulation factors in stored plasma components. There is no evidence that adding a requirement aimed at preserving Factor VIII content would result in an increase in the therapeutic value of plasma components. Furthermore, such a requirement could send clinicians the erroneous message that FFP can be used for Factor VIII replacement, when that component is clearly not the treatment of choice in case of Factor VIII deficiency.

AABB believes that given all of the competing priorities in transfusion medicine, there is no need to require changes in present methods of preparation of plasma for transfusion. Any changes to regulation should be based on evidence of beneficial clinical outcomes.

AABB believes that the time and effort spent in reducing inappropriate ordering of plasma transfusions by clinicians would produce greater benefits to patients than would result from focusing on changes in plasma preparation. A commitment of resources and research dollars to the plasma production area will detract from other issues that can better benefit clinical outcomes.

Thank you.

DR. ALLEN: Thank you. That certainly lays out one of the spectrum very well.

Any questions or clarification?

DR. DI MICHELE: I'm sorry, I just kind of walked in, in the middle of this. And maybe this has already been discussed, but is it not true that you have interchangeable pools of plasma for transfusion in recovered plasma?

MS. GREGORY: No, plasma for transfusion and recovered plasma are not intermixed.

DR. DI MICHELE: Don't you use FFP that's about to expire for recovered plasma?

MS. GREGORY: We would like to.

DR. DI MICHELE: But you don't?

MS. GREGORY: I believe after it is expired we can convert it to recovered plasma.

DR. DI MICHELE: Okay, so then recovered plasma does got into --

MS. GREGORY: Recovered plasma does go into pools for further manufacture.

DR. DI MICHELE: Including Factor VIII products?

MS. GREGORY: Most of the recovered plasma is not used for Factor VIII products. I'm not sure if any of it is.

DR. DI MICHELE: That's not what we understood from the FDA workshop. So, I just want to clarify that I think you can't make Factor VIII totally irrelevant as long as there are interchangeable pools of plasma for transfusion and recovered plasma that are going to go into Factor VIII plasma-derived concentrates.

MS. GREGORY: Well, plasma that is used for recovered plasma has to be manufactured to meet the specifications of the fractionator. So, whatever those requirements are, are what we would have to do. And if our fresh, frozen plasma did not meet those requirements, then we would not be able to convert FFP into recovered plasma for Factor VIII.

DR. DI MICHELE: All I'm just saying is that Factor VIII is not irrelevant.

MS. GREGORY: Factor VIII is not irrelevant if you are manufacturing a product that you need the Factor VIII for. Otherwise, for transfusion purposes, we believe it is irrelevant.

DR. DI MICHELE: And I would agree with you, and that was the nature of my question before to the presenter, that essentially we don't use plasma for transfusion generally for Factor VIII. Like I said, as long as those pools are interchangeable, it is used for Factor VIII, and I just wanted to make sure that that was clarified.

DR. EPSTEIN: Let me try to clarify the question for Dr. DiMichele. First of all, with regard to FFP, there is FFP made by apheresis, and there is FFP made from whole blood. FFP by apheresis cannot be used for fractionation until after it's outdate, because otherwise it would be source plasma. And that is a distinction that the industry would like us to remove, so that FFP, if not needed for transfusion, could then be sold at any time for further fractionation use.

However, your point remains valid, that because one day after 365, it could then be sold for fractionation, and it sometimes is, and it does contribute to the fractionation pool. So, the quality for Factor VIII does matter, and that Factor VIII is also made from that product.

Now, the recovered plasma that is made from whole blood collection is salable immediately for further use in fractionation, and indeed a very large proportion -- I don't have an exact figure -- but a very large proportion of the plasma made from the whole blood collection is immediate sold for further use in fractionation. So, once again, the clotting factor concentrations may matter.

Now, this subject was discussed extensively at the FDA workshop in September of last year, and the point of view of the industry is that although there may be different yields of Factor VIII in plasma provided for further fractionation made by a variety of means, that there is no apparent difference to the end products that result from the fractionation process.

Now, there is an arguable issue whether quality factors might exist, because if part of the issue of yield is really due to degradation, then we might be concerned about degradation by-products that could end up in final products. But that's an area where we begin to deal with unknowns.

All we can say is that given a diverse set of source materials, the fractionators know how to achieve target potency of the end product. And the industry argument is that the yield of that process is not an FDA concern. In other words, if they have to fractionate more plasma to get the same number of units of Factor VIII, as long as product quality is not compromised, why should we care? And the industry further argues that because of improved methods of improving the yield in fractionation, that there also has not been an issue of availability.

But the correct answer is that recovered plasma and outdated FFP made from apheresis collection do contribute to common pools to make all the fractionated, including coagulation factors.

DR. ALLEN: Any other questions?

DR. LEITMAN: Can I ask a question for the pediatricians that treat children with von Willebrand's disease. Is there a role in von Willebrand's disease for treatment with cryoprecipitate?

DR. MANNO: We don't use cryoprecipitate any more, we use the plasma-derived Factor VIII products that retain von Willebrand Factor activity.

DR. LEITMAN: The only reason I could think of to preserve optimum Factor VIII would be when you are making cryoprecipitate. But most of the cryo we use is not used for Factor VIII.

DR. DI MICHELE: Certainly in this country, yes, that's true. But on the other hand let's say we have patients with a very severe form of von Willebrand's disease who end up in a small hospital where there isn't one of the approved von Willebrand Factor concentrates. Cryo would be what we use, but I agree with Dr. Manno, it's only used in a very emergent situation, and is certainly not the product of choice.

MS. BAKER: I would also agree with that. And we have to remember that only about two-thirds of hemophilia patients -- and we are not even talking about von Willebrand's, which is much more prevalent -- are seen at the federally supported hemophilia treatment centers. So, there is this whole swath of von Willebrand's patients that are just out there at a variety of local hematologists and hospitals that have no affiliation with hemophilia treatment centers, small community hospitals that are taking care of von Willebrand's patients urgently, and do indeed use cryo.

And I have an anecdote. I'm not a treating physician. I'm not a nurse, but I got a call within the past three months from such a case up in Bakersfield in California, where a physician only had cryo available for a bleeding von Willebrand's patient. So, it is used out there.

DR. ALLEN: Okay, thank you.

Our third speaker is Mr. Joshua Penrod, Manager, Regulatory Policy of the Plasma Protein Therapeutics Association, and again speaking on review of standards for plasma products for transfusion.

MR. PENROD: My name is Josh Penrod. I'm Manager of Regulatory Policy from PPTA. I'm a salaried employee.

In other for Dr. Allen not to have to use the brevity hammer, Mr. Bult spoke this morning about who and what PPTA, so I can skip that introduction. And I will just give our brief statement as follows.

PPTA, on behalf of its member companies, supports the view of AABB. AABB has noted that the majority of the literature cited by the FDA in its briefing documents uses the measurement of Factor VIII as the standard for judging plasma quality. AABB has stated that, "Factor VIII is generally not clinically relevant to the therapeutic benefit of plasma components."

Recovered plasma prepared in blood centers using techniques and equipment similar to those used for the preparation of transfusible plasma components as a starting material for the manufacture of plasma therapies. At the August 31 and September 1, 2004 plasma standards workshop the plasma therapeutics industry repeatedly stated that Factor VIII is not the driver for plasma therapeutics manufacturing.

We agree with AABB that, "there is no evidence that adding a requirement aimed at preserving Factor VIII content would result in an increase in the therapeutic value of plasma components," and for our industry would not enhance the manufacture of plasma therapeutics.

We note that there are no current difficulties in the treatment areas specifically mentioned by the agency in this briefing document. While some of the questions raised in a review of the literature are interesting, and many questions remain unanswered, there lacks a science-based consensus on which regulatory action should be based. In the absence of scientific consensus or a public health concern, no agency action is needed.

PPTA echoes AABB's statement in terms of competing priorities, and would prefer to focus on areas of safety and efficacy that will have a measurable positive impact on patients.

Thank you.

DR. ALLEN: Thank you very much. Clarification questions for Mr. Penrod? Thank you.

It was very interesting to get the industry side of things.

Dr. Epstein?

DR. EPSTEIN: Yes, I wondered if I could be permitted to ask a clarifying question to Dr. Fitzpatrick?

I was very interested in your survey of current practices, and I wonder if you could comment. We believe that there is a fairly large amount of plasma which is sold for fractionation in Europe. And the European standard for plasma for further manufacturing is freezing at -30. And so, it would seem that in order to freeze at -30, centers must have -30 freezers.

So, I'm kind of struck by the absence of such in your report, and I wonder if you could give us any general sense for what proportion of recovered plasma is in fact being frozen at -30 in order to make it salable in Europe?

DR. FITZPATRICK: The 250,000 liters we sell annually are frozen at -30 at a subgroup of centers, and those are reflected in that percentage that had freezers capable of that.

DR. EPSTEIN: But then how does that factor as say a percent of volume, rather than a percent of centers? Or as a percent of all the plasma, how much of it is being frozen at -30? Because there may be a distortion in looking at it by number of centers.

DR. FITZPATRICK: Okay, 250,000 liters is approximately a little over a million donations, 1,200,000 donations. So, that gives you a framework at least for that.

DR. BIANCO: Just to help with -- Celso Bianco, America's Blood Centers, also a salaried employee. Just to help with a clarification requested by Dr. Epstein, there are two companies are the two only companies that manufacture products, derivatives from recovered plasma from members of ABC. One company is Zeta LB. The other company is Octapharma.

Zeta LB does not require freezing at -30. And they consume a substantial proportion of the plasma, I would say about 70 percent to 80 percent. And Octapharma that distributes products in Europe, follows the European regulations. Zeta LB only produces IV Ig that is distributed in the United States. I saw some of their members here. If there is a need, I'm sure they can help with that.

DR. WEINSTEIN: I wonder if I might also ask a clarifying question of Mike Fitzpatrick regarding the freezers. You mentioned that you had freezers at -18 degrees and freezers at -25, and so forth. Are you actually saying that the -18 freezers really have a set point of -20, or something?

DR. FITZPATRICK: That one column that says -25 capable means that they may have those set to store and maintain -18 or below, but they feel that those freezers have compressors capable of holding at -25, not -30. The jump from -18 to -30 usually takes a different type of compressor or a double cascade system to be reliable. So, that's where you had the column that said -25 capable.

And if you are holding at -18 or below, in order to keep within that temperature timeframe, and have time to react to an alarm when it goes off, the set point for the alarm is at least -20. It might be -21 or -22.

DR. WEINSTEIN: So, in fact these freezers that have the upper limit at -18, the reality of the situation is that they are mostly at -20, or some in fact are at -18? Maybe you can elaborate a little bit on that.

DR. FITZPATRICK: Some are probably at -18, especially in the hospital setting, or close to -18. And the other thing is depending on the type of freezer, if you have seen temperature charts from freezers, they cycle up and down as the compressor goes up and down. So, their highest temperature might be -19, their lowest temperature might be -21, and it cycles in between those temperatures.

DR. WEINSTEIN: There was another element of discussion here about having freezers just to freeze material at -30, and then to store it at -20 or -18. Was that raised in your analysis of what could be done, or what the cost would be for that scenario?

DR. FITZPATRICK: Yes, that was where the new equipment came in for freezing. If the freezing temperature had been -30 or below, then those sites that identified they would have needed new equipment, would have had to purchase that new equipment. And the survey discussed storage at -20, rather than -30 or -25. And our position at the workshop was that the storage at -25 might be feasible. Storage at -30 was not without a lot of additional expense.

DR. ALLEN: Dr. Fitzpatrick, let me just ask another question also. If I remember correctly, on your slide you had the -30 freezers in parenthesis and blast freeze. I assume that by that you mean one capable of very rapid freezing, which I assume means that there is special equipment, probably a different type of shelf structure and so on for circulation of ultra cold air?

DR. FITZPATRICK: Actually, the freezer that is most commonly used, and this is to comply our centers to sell to Octapharma to comply with their requirements for freezing within a certain time after collection or removal of the plasma from the red cells, circulates the super cooled liquid around an envelope. The plasma is put in the bladder. The super cooled liquid is circulated around that for heat transfer to make the rapid rate of freezing possible. So, they are very special freezers just for freezing. And then they are moved to storage after the freezing is complete.

DR. LEW: I did want to ask one question too. Do you feel comfortable that the people who sent in responses really represents nicely the broad spectrum of all the people enrolled in your organization? Or is it a possibility it might have been skewed, with some people very concerned about what it might do for their center in terms of cost? It's just kind of an open-ended question.

DR. FITZPATRICK: The respondents who are involved in contracts with Octapharma responded. That's not the bulk of the respondents. So, I feel that we got a fairly good representation of 68 percent of the collections of the organization. The hospital respondents, being a very small subset of that 3,300 is probably not as good a representation of hospitals, and you would have to do a further survey to know exactly what is going on with the hospitals.

DR. LEITMAN: I have a question for FDA. This was in our handouts. I don't think Dr. Weinstein covered it in his presentation. The difference between source plasma in the US and FFP in terms of long-term freezing conditions and shipping temperatures. Every now and then I have to deal with the two together, and there is a 2 degree difference, and it always befuddles me. Why is one -20, one -18? And then shipping for source is -5. Is there a reason for that?

DR. WEINSTEIN: The reason I think is again, this was found to be acceptable the purposes of manufacturing the product. In other words, the manufacturing process. I'm going to have to clarify here. If we are talking about the US standards for shipping for source plasma, it can exceed -20 for as much as 72 hours. It should never exceed -5 degrees.

And again, these conditions were found acceptable for them manufacturing purposes. But one can well question whether or not no deviations for freeze frozen plasma is appropriate, or could one think that perhaps those conditions could be revised, in fact allowing some degree of variation in the temperature fluctuation?

So, that's one of the questions I think that I'm proposing. If you look at the total picture of storage conditions, and conditions for preserving activity here. We find one of the slides showed that one could actually have some degree of fluctuation of material, and change perhaps the fibrinogen content, but not the Factor VIII content of your stored material in FFP. Now, is that so bad? I guess that's one of the things that one has to think about.

DR. LEITMAN: I think I was trying to make a plea for harmonization between US storage requirements for source plasma and FFP, and they were both -18, then aesthetically, without any data to suggest there is a difference in those, it would be easier to remember, and make more sense.

DR. KATZ: I have a question for somebody who is a clotter on the committee or in the audience. I was just looking at Dr. Szymanski's slide that describes the apparent yield at four different blood centers. And I would ask the question, is there any clinical implication of getting a unit or several from the ARC in Pennsylvania, and giving more than what you thought you were giving? Because that's not the apparent standard for these products or for this yield. Could somebody become pro-thrombotic for example?

DR. DI MICHELE: Probably, given the fact that you are starting from somewhere that is low. You are not replacing your entire plasma fraction with this product, probably not. Is there a reason to have a yield that that's high per unit of cryoprecipitate? In general, no.

I just want to say that about 10 years ago before the widespread use of recombinant there were several centers that had directed donor cryo programs, at which point family members would donate cryoprecipitate, and it would actually be given instead of a commercial concentrate.

At that time actually the donors were actually given DDAVP in an effort to raise their levels four fold to the 400 percent level, to increase the yield per volume. But short of that, there is probably no reason for a 400 percent level in cryoprecipitate.

It's unusual to get that kind of yield. It would have to be kind of a stress reaction, or some sort of activation phenomenon I think, but I don't know.

DR. ALLEN: Thank you. Dr. Weinstein.

DR. WEINSTEIN: This is another kind of general question perhaps to the treaters. Are there adverse events associated with any of these products that could be related to the time that they are in contact with cells, the time to freezing? In other words, apparently we don't use thawed plasma for infants for some reason. I don't know that reason. It wasn't exactly stated in a scientific manner. Is there some particular concern? Why not use it for that purpose?

But apparently there is a gut feeling that this isn't such a good idea. And just the general notion here, are there activated proteins, as was suggested in the graph that I showed of plasma that sits for 6 degrees for a long period of time, 34 days, or Liquid Plasma? Do people see reactions? We're not just talking about replacement. In other words, we're talking about adverse events associated with the use of these products.

MS. GREGORY: Let me answer about Liquid Plasma, because I'm not aware that anybody ever uses Liquid Plasma. It's there. It's in the definitions. We could do it. We haven't done it for years. Even when I trained as a medical technician we didn't do it.

DR. ALLEN: All right, it's almost 5:15 pm. We I think have got the basic information in hand. I'm sure there will be considerably more questions and discussion, but what I would like to do at this point is move back, and ask Dr. Weinstein to give the FDA perspective and questions for the committee. And then let's move into committee discussion and recommendations.

As Dr. Weinstein is coming up, I'll just sort of set the stage by saying there was the workshop last year. Mark, I guess you may want to give a 10 second summary on that. You indicated there were two large notebooks full of materials that came from that, but I guess not really a summary document that raised questions from it. It is I guess available on the Web site, is that correct?

DR. WEINSTEIN: Yes, it is available on the Web site, the slides and the summary of the meeting. Or at least they have been available. I believe they still are.

DR. ALLEN: I did not do due diligence and look at it in preparation for this, which I probably should have. But obviously the outcome from that workshop didn't give the full guidance that the FDA really is seeking, or else I assume they would have come to us with this question.

We have heard the industry response, which might be summarized in there doesn't seem to be a critical problem. It is going to be expensive if we make changes. We are not opposed to doing new research and learning more, but let's be very, very cautious about making regulatory changes.

Some of the committee members have raised other questions in terms of products for patient populations. Dr. Leitman I think very appropriately raised the same question that I had, which is if you got very similar types of products, why do you have a couple of degrees difference her and there? And how do we get to harmonization? And what's the impact of doing that?

And what does it really mean in terms of the question that was raised to the committee, the quality of the products? And I guess as I looked through the literature that was provided, the summaries and everything else, my head was reeling a little bit that there is a lot of data out there, but nobody has really, to my way of thinking, sat down and asked the critical questions, and the gone out recently at least, in a logical fashion, to try to answer them.

I suspect part of the problem is that there isn't a lot of money available. That was another question that was raised by the industry in terms of directed research. Is this the best place to put it? So, I think these are some of the questions to my way of thinking, that perhaps the committee would want to wrestle with a little bit as we try to provide some guidance to the FDA.

Agenda Item: II. Review of Standards for Plasma Products for Transfusion - C. FDA Perspective and Questions for the Committee

DR. WEINSTEIN: You summarized it very well. I think as we looked at the information that we had at the committee, and we look again at the workshop, and we look at what is in the CFR and the AB regulations, again, the issue is that right now many of these conditions are very flexible.

They allow a great deal of variation. There is potential for a large degree of fluctuation in Factor VIII, von Willebrand Factor, Factor V depending on the conditions for which these products are prepared, separated from whole blood, frozen, thawed, and so forth. So, that's one element.

Another element is potentially safety issues. And what is the relationship between some of these conditions and the reactions that patients have to plasma for transfusion products -- the hives, the fevers, the other elements here that are maybe not life-threatening, but do occur, are uncomfortable, perhaps could be avoided by a product that was prepared in an optimal way, whatever that optimal way might be.

The question about surrogate notion, the Factor VIII as being the most labile product. That is certainly what we look at. And it has been the point of study for many of these investigations that I have presented, because that was the product that was necessary in the eighties. People were looking for the highest yield to make the cryoprecipitate that was used to treat patients. And of course that has changed over time.

But the issue about whether Factor VIII is a surrogate for other labile proteins, things that we don't know about. We think we know a little bit about the TTP element perhaps, but I think that there are still questions about what the factors are that are necessary for the reversal of TTP.

So, we have presented some data that suggest you can make cryoprecipitate in a way that minimizes the level of von Willebrand Factor in cryosupernatant. And because these products are used for -- particularly FFP cannot be used for replacing things that we don't know about, and a broad spectrum of things, should we try to optimize the conditions for preserving as much of the native activities as we can? Those are the sorts of things that we are considering.

So, I can read again the questions to the committee. Please discuss the extent to which the available literature on plasma processing may support changes to improve the clinically relevant safety, purity and potency or consistency of various plasma products for use in transfusion such as the time to plasma separation, time from collection or separation to freezing, freezing rate and target temperature, storage temperature, allowed temperature variations during shipping and storage and cellular content.

Are additional scientific studies needed? Should we have some sort of task group to look at these issues and try to review the literature? What recommendations do you have for the next steps forward?

Agenda Item: II. Review of Committee Standards for Plasma Products for Transfusion - D. Committee Discussion and Recommendations

DR. ALLEN: Thank you. Questions or open for discussion.

If Dr. Fitzpatrick is still in the audience, just a quick question for him. Freeze frozen plasma needs to be, as I understand it, separated and frozen within 8 hours of collection. Am I correct on that standard? Given your small blood centers, a lot of mobile collection units and so on, I gathered from one of the slides that in fact the separation and freezing really is much more likely to happen 16-24 hours out, rather than in the first 8 hours. Is that correct?

DR. FITZPATRICK: If it's being made into recovered plasma, there are different requirements for recovered plasma. Some are very stringent, requiring freezing within a few hours of collection, 8 hours, 12 hours, and then 72 hours. But that is for recovered plasma, depending on what the manufacturer is going to do with it when he receives it.

So, we provide a number of liters based on the request of the manufacturer for production of IV Ig or albumin. And the longer period to freezing are primarily for IV Ig or albumin. The shorter periods are for recovered plasma that goes into manufacture of Factor concentrate.

For fresh frozen plasma, we meet the 8 hour requirement. So, the plasma is either made into fresh frozen plasma by meeting the time requirement, or it goes into those other products.

DR. ALLEN: Okay, I guess my question was really more one of the practicality. Given in particular the collection on mobile units and the need to transport, and that sort of thing, is the 8 hour standard not practical for a large percentage of what's actually collected?

MS. GREGORY: I think that's a reason why a number of places have gone to instead of making fresh frozen plasma, they make plasma frozen in 24 hours. And the point I was trying to make in my statement is that those two products are used interchangeably for transfusion purposes. And that's what these questions have to do, is with transfusible plasma, not so much the plasma that is being used for recovered plasma and for manufacture.

As a matter of fact, once we thaw the plasma out, we call it either fresh frozen plasma or plasma frozen within 24 hours. We leave that same name on it for 24 hours after it is thawed out. After that it becomes thawed plasma, and it's good for an additional four days I think. And they use these three products interchangeably.

DR. FITZPATRICK: For the European side of the house, since that has been brought in, just recall that most of the red cells there, they make a buffy coat platelet, and the red cells there are held at room temperature for up to 24 hours before the product is separated. So, it's an entirely different plasma product at that point.

DR. KATZ: The work flow varies at what blood center you are at, but in general we plan very carefully that we know how much FFP we need. And for example in my center we have several fixed sites within a short enough distance of the mother ship, that that's where our FFP comes from, so that we get around that process issue of 8 hours versus 9 hours, by planning production very carefully. So, that's our FFP. And the stuff that comes from someplace else goes into recovered plasma. And then one way or another, as we like to say in Iowa, everything but the oink gets used.

DR. DI MICHELE: Can I just ask a point of clarification from Mark then? Let me go back to this. Are we just talking about plasma? Are we just talking about FFP? Whatever it's called FP24, and thawed plasma, is that all we're talking about here? We are not talking about plasma that goes into recovered plasma? Many times we were talking about recovered plasma and source plasma, and we are having a different conversation. So, maybe let's just clarify that.

DR. WEINSTEIN: The focus of this particular session here is on transfusible plasma.

DR. DI MICHELE: Okay, so if that's the case, and if going back to one of my earlier statements then, it sounds like the most labile of all the factors that we need to transfuse for per se, not that goes into manufacturing is Factor V, that we know of.

DR. WEINSTEIN: Right, and we have loss of Factor V activity also that is reflected.

DR. DI MICHELE: So, is not the starting point then -- I think to most people who use these products, I don't know that -- I hope I have never used thawed plasma. I don't know if anyone has ever given me thawed plasma to use when I have order FFP, but I hope to God I'm not using thawed plasma.

DR. WEINSTEIN: Again, this is not a licensed product either. Thawed plasma is not a licensed product of the FDA.

DR. DI MICHELE: But based on what Kay is saying, if I'm ordering FFP, a blood bank may give me thawed plasma?

MS. GREGORY: You should be able to tell that by looking at the label, because it will be labeled thawed plasma.

DR. DI MICHELE: I never look at the labels, but I guess I better start reading labels. It's kind of like the grocery store. Okay, so maybe one of the ways to start might be to really understand, and to further characterize if they really haven't been to any kind of satisfaction with respect to the clotting factors for which we use them for, to characterize these different products so that we understand what we're giving when we give these products, and really understand if these are interchangeable products.

That would almost sound like based on current processes, it sounds like that may be the first place to start.

DR. QUIROLO: My understanding is about 30 percent of the time people who give these products, give the wrong product for the wrong reason, maybe even more. With AABB thought maybe some of this research money should go into figuring out how to teach doctors to give the products the proper way. In our hospital it took months to get the gastroenterologists not to give two units of FFP before the did a liver biopsy, or one unit to a 200 pound person to do a liver biopsy or something like that.

So, I think the issue is there are so many products, that even the people on the committee here don't really know which one is which, and which one they are giving when they are in their hospital, which I think is a big problem. There are too many products, and too many ways that they are made. In our hospital we don't even have FFP anymore. It's all the F24 or whatever it's called.

DR. LEITMAN: If there was one product you were going to drop, it would have to be Liquid Plasma, since 99 percent of all blood -- 100 percent is manufactured in packed red cells with additive solution, there is no whole blood on the shelf at the end of the storage period of the unit any longer. So, you can't make what is defined as Liquid Plasma. So, it's recognized, it's licensed, but it's never made.

I can think of a setting where we may have used it, and I think we did use thawed plasma, which is if a patient for whom you thought plasma, doesn't use it, and it's been thawed for longer than 24 hours, and at exactly the same time you happen to have a patient receiving plasma exchange to treat TTP, that would be a very good use of thawed plasma.

DR. LEW: Just as a comment though, to mention although I agree, I think physician practices can improve in a lot of areas that could save a lot of health dollars. But the key is will we use it appropriately if we have the money? Just as we know you can save a lot of lives if we developed a vaccine for malaria, but is the money going to be there?

So, I guess the question that we are thrown at is not if we had X number of dollars, we can use it for one or the other? My understanding is FDA is asking is there something that we need to ask, such as the question we have these products. Should we know how good each product is for what we will be using it for? That seems to me a legitimate question. Although, I agree for public health reasons, boy, our dollars can go further if we taught physicians.

DR. DI MICHELE: But I don't exactly know what dollars people are talking about in terms of teaching physicians. I mean physicians should be taught as part of the -- that's the mandate of their residency training programs. The hospital blood banks should be doing this teaching. So, I don't exactly know what public dollars should really be going into educating people. This really should be part of the training mandate for both hematologists and any practicing physician. Don't you agree, Kay?

DR. ALLEN: Well, certainly one of the issues though is changing standards, and how you get physicians in practice at hospitals across the country to do that. And I'm not arguing that that's an FDA function. I don't think it is. But it is a transfusion medicine function. I remember very much as I was going through my training program, this is right at the time that there was a large scale move away from fresh whole blood to component therapy. And our hospital blood bank put out guidelines, had seminars, and then put regulations into practice.

We had to literally provide a written justification for why we wanted to fresh whole blood. Well, my early training had been on fresh whole blood, and suddenly I was having to come up with -- it was much easier to go with packed red cells. So, this issue of continuing training is an important one, but it's not an FDA function in my view.

DR. SCHREIBER: Maybe because it's late in the day and I'm getting a little confused, but I thought what we were supposed to be talking about were standards and storage conditions relating to the plasma. And the discussion about use and teaching physicians is very interesting, but when I look through and try to get from the literature, the points relating to relevant safety, purity, potency, or consistency of various plasma products used for transfusion, I couldn't find any of those parameters in any of the pack of literature that we got.

So, from my standpoint, I would have a very tough time making any intelligent comments about what happens if you store plasma under different conditions. I can find that we have a number of measures, whether it's Factor VIII or protein or whatever, and I can see that some of them increase or decrease marginally over time. But that doesn't tell me anything about the clinical relevance of those changes, the potency of the product.

And I think that from my standpoint, there is just not enough known about what the total impact for patient care would be in terms of changing the requirements. In terms of the storage, if you are just measuring the parameters of Factor VIII or V or plasmagen or something else, from the literature that we got to review, I didn't see any significant changes.

DR. ALLEN: Thanks for those comments, George. I agree with you. And I think my sense is the FDA still today is dealing from a scientific basis, with a lot of older data that has been accumulated for whatever reasons. And there really hasn't been a thoughtful approach.

The other issue that the technology of collection continues to change rapidly. And I think that that perhaps needs to be factored in also. There clearly are many areas where there just needs to be some basic -- it's not even that complicated research studies. There needs to be funding for it. It needs to be carried out in a satisfactory way that mimics real life collection situations in blood collection centers across the country, and that deals with the practicalities or the practical limits that are placed on blood collection centers.

Again, going back to Susan's point, a little harmonization and simplification would be a good thing. And I guess my overall summary suggestion to the FDA would be that perhaps I suspect they don't have staff time to really put into this, given everything else. But perhaps there could be some sort of an advisory panel that would be empowered over a period of a year to work through some of this, and come up with a very targeted research agenda that might be pursued to answer the most specific questions.

I am sensitive to the industry's requirements. Don't just make arbitrary decisions. I think they do need to be science-based decisions. I'm not sure if you've got limited resources, that this is going to dramatically improve the quality of patient care, but I wouldn't rule that out if in fact we were able to collect better science.

But I think as I look through the papers, and listen to the discussion, there are spotty data all over the board without a really comprehensive -- there isn't a comprehensive set of data that allow determinations of the kind the FDA I think is trying to say we would like to make.

DR. KLEIN: Jim, my question is what is the clinical problem now that we are trying to address? If it's the side effects, the adverse events, probably the big one is TRALI, and we already know that that's the number one issue that Heart, Lung, and Blood is funding.

No one, I don't think, is going to put in an awful lot of effort into hives and into the febrile reactions we see, especially since there aren't any really good ideas about how changing the manufacture might change those adverse events, and to prove it would be a really large undertaking. It's certainly not at the top of your list.

If we knew precisely what we needed in this product, then I think we could certainly say, well, here is the level of Factor V or Factor VII, and now we need to determine what the best conditions are to both produce it and to retain it over time. But in many ways this is sort of like chicken soup. At least most of the plasma that is used in the United States is used for things that are very hard to measure outcomes on.

And whether you want to spend a lot of time standardizing Factor VIII or fibrinogen in order to get a product that is more standardized, but may not have any difference that you can measure in terms of clinical outcome I think would be a waste of time and effort myself.

I would also point out what I guess most everybody here knows, and that is that the amount that you get from an individual donor in clotting factor certainly, and most every other protein is quite variable, from 50 percent to 150 percent of so-called normal. So, you are already starting out with something that is quite variable. And trying to standardize it in any way, rather than giving a minimal dosage, as we used to do with Factor VIII and fibrinogen, or trying to determine a minimal amount of loss in the process, again would be doomed to clinical failure.

So, I'm not sure what the question is clinically that I could help you with.

DR. ALLEN: Dr. Weinstein or Dr. Epstein, do you want to try to respond back to guide our further discussion?

DR. EPSTEIN: Well, the larger envelope here is that we had started out over a several year period with a discussion on developing licensing standards for recovered plasma. And that led to the issue of well, how much like source plasma should recovered plasma be in order for it to be a suitable material for further fractionation use.

And then we get into the question of well, what is recovered plasma? And we discover that sometimes it's FFP-like, and sometimes it's not FFP-like. And then we look at the question of well, if we are going to put forward standards, and we're going to put them in regulations, well, they ought to be science-based. So, what's the scientific evidence on plasma qualities related to manufacturing methods?

And there what you come across is the literature that Mark just reviewed, which is that you have some data. The data are very helpful if you look at certain parameters that happen to be studied at certain points in time. And there is a big gap between knowing what happens to Factor VIII or Factor V, and knowing whether it matters clinically for any given use indication.

So, where we have kind of ended up is we don't have really a good handle on what the science is telling us as far as creating product standards. And yet, we are not happy with the product standards, because they are really not very helpful. As Dr. Weinstein showed, under the existing regulations, you are not making the same product in different places at different places at different times in different ways.

Now, we could take a step back and say, well, none of that matters. And a little bit I'm kind of hearing that, or at least there is no obvious problem we are trying to solve. But on the other hand, it is a very disquieting situation to think that we are regulating products according to standards, but the standards are not actually meaningful in terms of creating product consistency, and even less meaningful in terms of their clinical importance.

And so, what we are trying to do is figure out, well, where do we go with all of this? Do we just take a snooze for another 10 or 30 years? Or do we try to push forward and develop the data that would help us have a more rationale scheme which would include a more harmonized scheme?

Why exactly should we have -18 storage for FFP, and -20 storage for source plasma? Source plasma should be frozen immediately, but we don't define immediately. It should be placed at -20, but we don't have a parameter for rate of freezing. Does rate of freezing matter? Does temperature of storage matter? Does temperature of freezing matter? Does time in contact with cells at room temperature versus 1-4 degrees matter?

So, what we are seeking is a path forward. It's not that we are interested in jumping tomorrow to create regulatory standards that are not rational or don't have a clear relationship to the available scientific data. But looking toward the future, we think this is a very unsatisfactory situation from the standpoint of standardizing products, and harmonizing the existing disparities. And of course we only want to do it in a sensible way, but how do we get there?

DR. KLEIN: Jay, I think if I could just take a shot at that, if what you wanted, and I think it's something that could be helpful, is a definition of a process to standardize fresh frozen plasma, I think certainly you could get a task force together, as they did for Bethesda units many years ago, and say this is the way we are going to do it.

We are going to collect it after so many hours. We are going to freeze it in this way. We're going to store it in this way. We're going to thaw it in this way. And then we're going to analyze it for 20 different things, and we're going to say this is what it's going to take.

And maybe that's a good thing. You would have a relatively standardized product that would become known as fresh frozen plasma by a variety of measurements and by a process. Whether that would be relevant clinically, it probably wouldn't be, but perhaps that's not as important. I don't know. You would have a standardized product that perhaps then you could do trials with if you so desired.

But I guess coming from the other direction, to say let's figure out what the outcome of important measures are, and then build our product around that, I think it would be much more difficult to do.

DR. BIANCO: I want to help Dr. Epstein, suggesting that there are areas that can be standardized. One is CGMP. This is a chicken soup, but it's a chicken soup like Dr. Klein said, that has to be prepared in a clean kitchen, under a closed system, under certain volumes, certain rules that have to be established.

The second part of the CGMP is the label. That is, if the unit label contains all those characteristics, if Octapharma wants plasma that was frozen within 8 hours with an AEO type label on it, because they still manufacture and distribute solvent detergent, treated plasma in Europe, that would be on the label, the same way a 72 hour plasma that is going only for the immunoglobulin would be in the label.

Those types of standardization through CGMP to definitions are easy. Pharmacologically, I doubt I think that everybody raised it very clearly, that we will find a set of points or activities or measurements that will define chicken soup.

DR. KATZ: And I really believe very strongly that the requirements for recovered plasma in large part should be between those of us who are recovering plasma, and those who are buying our plasma and making their products. And I'm struggling for the relationship of transfusible plasma to those requirements, although the harmonization aspect, I think that's a different discussion. But ZLB likes the plasma that I give them, and it's not FFP.

DR. DI MICHELE: I would just like to second what Dr. Klein said. Again, going back to the fact that we are not discussing recovered plasma at this point, we are discussing plasma for transfusion, because I get very confused all the time, but if we are discussing plasma for transfusion, then I would agree that it is a little confusing to have three products that are currently used interchangeably.

And I think that if we look at the current guidelines for what we use plasma, we use them for things other than these guidelines. But let's say we just start taking the guidelines for what we use plasma for, and we say okay, what are the minimum requirements that we need here? What's the most labile factor? Because you are not going to have plasma that is just designated for Factor V replacement, or situations of liver disease when you need to replace Factor V.

So, what is the minimum requirement? And then what are the minimum conditions that it's going to take to actually produce that product? If that's the case, maybe we don't need plasma for regular transfusion that's ever frozen within 8 hours. Maybe everybody can just freeze it within 24 hours, and let's be done with it, and store it at -18 if that's what the industry is currently doing, that would be okay.

So, I really do believe, like what I said before, that we really have to start with looking at all the products we have, deciding what the minimum requirements for the transfusion requirements that we have now, the minimum content requirements for that plasma would be for safety, and for presumed efficacy.

Because I agree with you, there has never been an efficacy study that has really been done, but let's say in other words just at least to tell us what we are giving is what we expect to be giving, based on what we have been taught. And if not, what are the ranges, so we know what we are giving. And then decide, should thawed plasma ever be given? I don't know. Maybe eventually we'll do away with both fresh frozen plasma and thawed plasma and we'll just use the 24 hour stuff.

But I think that's where we should start, if that's where we are going to start. And then as you said, I think that creates a standard product that then we have for transfusion. And then we can kind of go from there in terms of if we want to do any efficacy studies, or use it for other reasons, et cetera. At least we have a well characterized product.

Does that make sense?

DR. LEW: And I think to add to that, I want to go back to the study that looked at Philadelphia's center. And I guess to the hematologists, it is kind of striking that they had so much better recovery. Has any study looked at the efficacy of that? Because we are all trying to also limit blood products if we can. And I thought if you had twice the yield, that was really striking, because of simple recovery processes. Maybe that's a good thing, and we could use less product. Has that been studied? I don't know.

DR. ALLEN: Okay, it's not quite 5:50 pm, but it's getting close. Other critical comments or thoughts for the FDA on this issue? I think there has been a lot of discussion. It's been very general, Jay, and I apologize for that. On the other hand, I think the committee members have struggled with how to provide some sort of guidance.

DR. EPSTEIN: I think the general discussion accurately reflects the current state of knowledge, and I think that there certainly are few take home messages, at least that I've heard. One of which is that some standardization in this area would be desirable. And the second is that some kind of task force of experts could attempt to develop a consensus position on a desirable standardized product. That may just be as far as we are going to go.

I think I also heard that some clarification in labeling, so that the user has a better idea of what the product really is would be simplifying in their own right. And that that same schema could apply to the beside user, as well as to the fractionator. If there are other core points I missed, I don't mind hearing them again.

DR. KLEIN: In some ways, Jay, maybe this is more like an herbal remedy, where the first place you would like to start, is you would like to know how much of this stuff is in the bottle. And maybe we don't know that yet. So, you can't even do a study with your herbal, because you haven't got a clue as to what's in there.

DR. ALLEN: Yes, but the standard with chicken soup is how it tastes.

MR. ROBINSON: Richard Robinson, American Red Cross. I'm just an employee of the American Red Cross. I'm not a clinician.

Currently, we have transfusible plasma collected in the United States under one set of conditions, transfusible plasma collected in Europe under another set of conditions. And I've not heard that there is any sort of increment of safety or efficacy between the two. Likewise with the plasma derivatives that are collected and manufactured in the United States, as opposed to those in Europe, I haven't heard any comparative claim of superiority one over the other. So, I don't know that there is a compelling reason to look at more stringent standards than what we have.

But what I haven't heard is the business impact. And I think that is one area where additional information is needed. What's going to be the impact to the blood centers to retool if we go to more stringent standards?

DR. ALLEN: Thank you.

DR. LEW: If I could just address that. I don't think anybody on the committee or FDA is saying we want to make any changes today. But if you don't look, you don't find, and that's true in medicine and anything else. So yes, there has been no great comparisons between how they do it in Europe versus the United States, and how good those products are. But that doesn't mean you shouldn't be looking.

Because now I think it's reasonable to think if FDA is responsible for regulation, they would like to be able to do it based on science, not because no one has complained yet.

DR. ALLEN: I think that's a very good point. And certainly, blood collection centers across the United States are getting more pheresis equipment in because of the economics, and it enables them to selectively collect the components that they want. Increasingly, leukoreduction is becoming a standard. That's a costly thing, but from a clinical perspective it's believed to be worthwhile. Other changes are being made.

If there is justification on a scientific and a clinical basis for doing this, if we can reach harmonization, I think we should not shy away from change. And probably in the long-term it will definitely be better. But I understand Jay's problem in terms of trying to help line out an agenda that is appropriate and putting a process in place.

DR. KLEIN: I just want to clarify, I didn't say anything about being more stringent or less stringent. I made an argument for standardization so that you know what's in the bag. That may be more stringent. It may be less stringent. It may be storing at 12 degrees, for all I know. But at least you would know what you had, and then be able to test the clinical outcomes with it.

DR. ALLEN: Thank you. All right, any other questions? Okay, a quick question about tomorrow morning. We are scheduled to convene at 8:30 am tomorrow. We then have a 12:30 pm lunch, and then one more hour of discussion. I have had a couple of people ask about would there be some way of perhaps starting at 8:00 am, and then running through our business day and having lunch as an option after we adjourn, assuming we get through all of our business.

Is there a problem with reconvening at 8:00 am, Jay, FDA staff, or Bill?

DR. FREAS: There is no serious problem at attempting to begin at 8:00 am. I'm assuming that all our speakers will be here and everything will be on track, and if all the committee members are here, we will do our best to begin. We may not be able to contact them. I just realized that two of the speakers are non-FDA speakers. And will not be able to contact them tonight.

DR. ALLEN: Well, what I would suggest is let's plan to reconvene very promptly at 8:30 am, get underway at 8:30 am tomorrow morning, and make the most of the time available to us, and see how it goes.

All right, we are in recess.

[Whereupon, the meeting was recessed at 5:55 pm, to reconvene the following day, Friday, March 18, 2005, at 8:30 am.]