DEPARTMENT OF HEALTH AND HUMAN SERVICES
FOOD AND DRUG ADMINISTRATION
CENTER FOR BIOLOGICS EVALUATION AND RESEARCH
This transcript has not been edited or corrected, but appears as received from the commercial transcribing service. Accordingly the Food and Drug Administration makes no representation as to its accuracy.
BLOOD PRODUCTS ADVISORY COMMITTEE
Friday, March 14, 2003
620 Perry Parkway
Kenrad E. Nelson, M.D., Chairman
Linda A. Smallwood, Ph.D. Executive Secretary
James R. Allen, M.D.
Kenneth David, Jr., M.D.
G. Michael Fitzpatrick, Ph.D.
Jonathan C. Goldsmith, M.D.
Harvey G. Klein, M.D.
Suman Laal, Ph.D.
Judy F. Lew, M.D.
Paul J. Schmidt, M.D.
Robert J. Fallat, M.D.
NON-VOTING INDUSTRY REPRESENTATIVE
D. Michael Strong, Ph.D., BCLD (ABB)
Mark Brecher, M.D.
Michael P. Busch, M.D., Ph.D.
Lyle R. Petersen, M.D., M.P.H.
Ruby N.I. Pietersz, M.D., Ph.D.
Edward Snyder, M.D.
C O N T E N T S
Anticoagulants, Irradiation and Freezing
of Blood Components,
Judy Ciaraldi, M.T., OBRR 7
Bar Code Label Requirement for Human
Drug Products and Blood,
Richard Lewis, Ph.D., OBRR 20
Open Public Hearing:
Dean Elfath, M.D., Baxter Healthcare
Open Committee Discussion:
Discussion on Extension of the Storage
Period for Pooled Platelets,
Jaro Vostal, M.D., Ph.D., OBRR 37
Clinical Performance of Pre-Storage
Pooled Platelet Products,
Edward Snyder, M.D., Yale University 45
European Experience with Extended
Storage of Platelet Pools,
Ruby Pietersz, M.D., Ph.D. 96
Bacterial Detection in Platelet
Products, Mark Brecher, M.D.,
University of North Carolina 114
Open Public Hearing:
Louis Katz, M.D., America's Blood Centers 155
Kay Gregory, MS, MT, AABB 158
Questions for the Committee 166
Update on Particulates in Blood Bags:
Introduction, Richard Lewis, Ph.D., OBRR 206
Discovery, ADR Investigation, Conditions
of Collections Investigation, Peter Page,
M.D., ARC 211
Chronology and Field Overview,
Jerome Davis, Ph.D. 239
Testing, FDA Findings,
Betsy Poindexter, FDA 250
C O N T E N T S (Continued)
Steve Binion, Ph.D., Baxter 261
Follow-up ADR Monitoring:
Clinical Studies, Sharyn Orton,
Ph.D., OBRR 275
Centers for Disease Control and Georgia
State Division of Public Health,
Matthew Kuehnert, M.D.,CDC 284
Summary, Richard Lewis,
Ph.D., OBRR 294
P R O C E E D I N G S
DR. SMALLWOOD: Welcome to the second day of the Blood Products Advisory Committee meeting. I am Linda Smallwood, the Executive Secretary of the Blood Products Advisory Committee. Yesterday I read the conflict of interest statement that applies to this meeting. If there are any declarations regarding conflict of interest that need to be made for anyone participating in this meeting, please speak at this time. If not, then at this time I would like to ask Catherine McComus to come forward and just make a brief request regarding a survey that is being done regarding FDA advisory committees.
MS. MCCOMUS: Thank you, Dr. Smallwood. Good morning. For those of you that were here yesterday, I apologize for repeating my plea. For those of you that are new today, I will be brief and ask you for your assistance to complete a survey. I have extra copies, in case you didn't get one yesterday, sitting on the registration desk. It is called conflict of interest and FDA advisory committees. I am asking your assistance to fill it out in regards to your opinions and what you know about how the FDA monitors the real or potential conflicts of interest of its advisory committee members.
The survey should take about 15 minutes to complete. If you have a chance to do it before you leave the meeting there is a box where you can deposit it. If you don't have a chance, there is a postage-paid envelope. You can drop it in the mail and return it. If the advisory committee members have misplaced the copy of their survey, I have extras of those too.
I will be here today. If you have any questions, I would be happy to talk to you more about the questionnaire and the survey. Again, I thank you very much for your assistance. Your opinions are very important and they will help to increase the validity of the study. Thank you very much.
DR. SMALLWOOD: As you may note from the agenda, we have another full day and we will be timing speakers so that we can try to meet our schedule as planned.
Today we will have ten advisory committee members that are eligible to vote. Dr. Liana Harvath is not with us today. Also, I would like to let you know that there is an electronic pointer here that is very small, like a pin, so when you finish making your presentation do not put it in your pocket. We will check.
At this time I would like to turn the proceedings of the meeting over to the Chairman, Dr. Kenrad Nelson.
DR. NELSON: Thank you, Dr. Smallwood. The first item on the agenda is a committee update on anticoagulants, irradiation and freezing of blood components, by Judy Ciaraldi.
MS. CIARALDI: Thank you very much. Good morning, everybody.
Today I am going to talk about anticoagulants, irradiation and freezing of blood components, a problem that we recently discovered and I think, at least for the moment, have resolved. I am going to describe how this issue came to light; give you some background information to explain the scope of the problem; let you in on what we discussed to develop our current thinking and then go over our current thinking.
We identified this problem during a routine review of a licensed blood center's labels. This licensed blood center is already approved to collect red blood cells by automated apheresis collection, and they are also already approved to irradiate blood components. They are currently using the codabar labels and wanted to switch to ISBT-128 labels so they sent in a whole stack of labels.
Among that stack was a label for irradiated apheresis red blood cells in ACDA/AS3. It is our custom during label reviews to check either in the codabar or ISBT catalogs to ensure that the firms are using the correct label code for the specific product. Initially we couldn't find an ISBT label code for this particular product. Upon further investigation and discussion, we determined that ACDA/AS3 units had not been evaluated for safety and efficacy after irradiation and after freezing. ACDA/AS3 is a novel anticoagulant solution that is used for automated red blood cell collection.
The outcome of the review was that we told the blood center that we could not approve their ISBT label for this particular product and that they should discontinue making irradiated red blood cells out of AS3 products. We also shared our concerns with the ABB standards committee and word spread through the blood community and there were a lot of concerns raised, as you can imagine. The AABB wrote a letter to Dr. Jesse Goodman asking for a clarification of this issue.
To give you a little background on what these products are used for, they have specific uses. Irradiated products are used to prevent graft versus host disease in certain transfusion patients. Frozen red blood cells are used to store rare cells up to ten years and they can also be used to create a reserve red cell inventory. Red blood cells can be irradiated only, frozen only or both irradiated and frozen.
We are concerned about the safety and efficacy of these products because we know that both processes damage the red cells, making them more prone to hemolysis and more rapidly cleared from the circulation during transfusion.
There are six anticoagulant preservative solutions that are approved for manual whole blood collections. They are listed on this slide. Of these, none are specifically approved to undergo irradiation and only two are specifically approved for freezing, the CPD and the CPDA-1.
There are four devices cleared for automated red cell collection and there are four anticoagulant preservative solutions that are approved to be used on these machines. The CPDA-1 and the CP2D/AS-3 are approved to be used on the Hemanatics MCS plus LN-8150 automated red cell collection device. The ACD-A/AS-1 is approved to be used on the Baxter Amicus and Baxter ALYX devices, and the ACD-A/AS-3 is approved to be used on the Gambro TRIMA device, and this is the one that was used at that licensed blood center. Of these, none is specifically approved to undergo irradiation and only one is approved to undergo freezing, the same one that was approved for the manual whole blood collection.
The more we dug into our files, the more we realized that this was a bigger issue, bigger than just novel anticoagulants and bigger than just the red cells collected on the automated collection devices.
These procedures have been in blood centers for some time. They have been doing these procedures for a while. To give you some thought about that or some information, we started approving anticoagulants back in 1959 with the approval of the ACD-A anticoagulant. The CPD and CPDA-1 anticoagulants were approved in the '70s. In the '80s we started approving anticoagulants that had additive solution for extended shelf life. The first one, ACD-1, was approved in 1983. In 1987 CPD and CPDA-1 were approved for frozen red cell storage up to ten years.
I think the blood banks will tell you that even though I have this listed as 1987 for ten-year storage, the blood banks were freezing red cells as early as the '70s with the anticoagulants that were approved at that time for product collection. That storage time was three years and that is all that was approved. Unfortunately, I couldn't find the specific approval date in our files. It goes back further than our electronic records.
The first irradiator was cleared by CDRH in 1991, and in 1997 CBER approved or cleared the first automated red cell collection device. There are currently 661 facilities that irradiate blood components and 233 facilities that freeze red blood cells. I tried to get a number of how many units have been frozen and irradiated, and there are a lot, and that data are not readily available at this time.
We spent the last few months discussing this issue. The discussions included officials from the Office of Blood Research and Review, specifically the Division of Blood Applications and the Division of Hematology. They also include officials from the Office of Compliance and Biologics Quality and lawyers from the Office of Chief Counsel.
We performed a risk-benefit analysis and considered the following issues: First, historically DBA has always reviewed irradiation as a process without regard to specific anticoagulants. This is a little before my time when we started reviewing these, but I think part of this was because when CDRH approved or cleared these devices the approval letter and the operator's manual did not place any restrictions on what anticoagulants can and cannot be used with the procedure.
In addition, there is a medical need for irradiated products to prevent graft versus host disease. Right now, for transfusion recipients that are at risk for this, it is the only procedure or process that will do this.
In 1993 and 1987 we published memoranda dealing with irradiation and freezing respectively. Neither of the memoranda mentioned any restrictions on anticoagulants. In fact, anticoagulants are not mentioned at all in the memoranda. Both processes are in wide use by many, many blood centers. In spite of that, we haven't received any reports of any adverse events.
But we are aware that the nation's blood supply is currently at a critically low level, and we know what the impact of asking for these products to be withdrawn or discontinued would be and that was certainly included in our considerations, and we are still concerned about the lack of safety and efficacy studies that have been submitted to CDER for these products.
While there are no specific submission approvals sent to FDA to review for irradiating and the variety of anticoagulants, we do have some information. Safety and efficacy data on irradiated products for some of the anticoagulant solutions were discussed at a '92 and '94 BPAC and we have those transcripts available. We reviewed the transcripts; we reviewed the literature. Based on that information and the considerations in the previous slide, it is our current thinking that blood centers may continue to irradiate products collected in approved anticoagulant solutions. What I mean by approved is approved for product collection, currently approved. FDA will seek data to support specific licensure of irradiated products in any new anticoagulants that are submitted to FDA to approval.
We consider it appropriate for blood centers to only freeze red cells in anticoagulants that are approved for this use. But because of the lack of adverse events reported during the long time that freezing has been performed in blood banks, and in the best interests of maintaining an adequate blood supply, blood centers may continue to freeze products in anticoagulants that are not specifically approved for this use.
Because we are concerned about the safety and efficacy of these products, we will seek data to support specific licensure of the anticoagulants used to prepare red blood cells, definitely on any new anticoagulants that come around the bend or are submitted to the FDA, but we would like to see if we can get some data on the ones that are currently approved for product collection.
To summarize, red blood cells are currently being irradiated and/or frozen under conditions not approved by FDA. We are concerned about the safety and efficacy of irradiated and frozen products. We are also concerned about the potential impact on the blood supply if we require these products to be withdrawn or for the manufacture of these products to be discontinued. Therefore, the current practices may continue with the anticoagulants that are approved for product red blood cell collection and storage.
We hope to resolve our safety and efficacy concerns quickly. Some of the automated red cell collection device manufacturers are currently sponsoring studies for the anticoagulants approved for their devices to their products to undergo irradiation and freezing. When data are available we will consider publishing a guidance.
To close the story on this particular blood center that brought this problem to light, we called them back and told them that we would approve their ISBT label for their product and that they could continue to irradiate that particular product. Needless to say, they were very happy to get that news. We also prepared a response to the AABB letter to Dr. Goodman, was sent to them on February 27.
On March 5th we published the contents of that letter in an information sheet on our CBER web site. The title of that information sheet is FDA current thinking on irradiating and/or freezing blood components collected and stored in anticoagulant/preservative solutions not specifically approved for such use.
As I was writing this talk as well as that information sheet, it struck me that the more complicated the issue gets, the longer the titles get of our documents. Thank you very much.
DR. NELSON: Thank you, Judy. Yes?
DR. FITZPATRICK: Judy, just for the record, I don't have a financial conflict of interest but I represent a group that has had a license application in for six years and probably has the largest repository of frozen red blood cells of any other organization.
I just wanted to clarify two things. On the first slide, indications for use and concerns, you say both processes damage red blood cells and that they are more rapidly cleared from circulation after transfusion. In fact, frozen red cells have to have a normal circulation after transfusion to be approved. So, they are not more rapidly cleared, although we know irradiated red cells are more rapidly cleared. So, I wanted to clarify that.
On the question of your two current thinking slides, I just want to make sure that we are clear. For irradiated products any anticoagulant that has been approved for collection of red cells can currently be used to be irradiated.
MS. CIARALDI: That is correct.
DR. FITZPATRICK: And continue to be a licensed product.
MS. CIARALDI: That is correct.
DR. FITZPATRICK: For frozen blood products, your statement, although worded the same, means something different. You are saying that only CPD and CPDA-1 are currently approved for freezing. Does that mean that those are the only products that can be licensed products as a frozen product?
MS. CIARALDI: It is my understanding--I wasn't present at the very last discussion and, in fact, I think I was inspecting your facilities, and I hope Alan Williams will come up and fill in the gaps here--the discussion centered around the need to evaluate any new anticoagulants that come around, and I know that wasn't your question. But we were hoping that blood centers would submit data to us, retrospective data on the products that are currently on the market, currently being frozen, and they could provide some data to allow us to license these products officially. Alan, could you come up and clear that up?
DR. WILLIAMS: It is a little bit of a quandary, but the answer to the question is, yes, they would continue to be licensed products but we do seek data to review the situation and will make an additional determination in the future but they would be considered licensed products, as they have been.
DR. FITZPATRICK: A product collected in AS-3 could be labeled as a licensed product and frozen
DR. WILLIAMS: That is correct.
DR. SIEGEL: Yes, I think the nuance here is that because we have specifically approved anticoagulants for freezing we are encouraging the use of the specifically approved anticoagulants, but we are not taking a compliance attitude toward the use of approved anticoagulants that are not specifically approved for freezing. We are treating it as a policy matter and an ongoing scientific issue but it is not violative.
DR. FITZPATRICK: If you received a request for a new label for a product using a different anticoagulant, would that be held in abeyance for data or would that be approved at this point?
DR. SIEGEL: It would be approved as long as it wasn't a novel anticoagulant.
DR. FITZPATRICK: Thank you.
DR. NELSON: Thank you. Yes?
DR. STRONG: I just wanted to point out that I also don't have any conflict of interest. I used to wear a blue suit similar to his. But the Navy has had data probably dating into the '60s. Bob Valeri had data on 20-year stored red cells and has done the experimental work in at least a variety of anticoagulants. So, if it is data that we need, they probably have the most of anybody in the world.
DR. FITZPATRICK: Just to clarify, that license submission that we have been working on for almost seven years now is for licensure for 20 years for frozen blood, and the FDA has a lot of data on that. It is on units collected in CPDA-1 and CPDA does not include units collected in other anticoagulants at this point in time. But there is a huge amount of data available on those two anticoagulants.
DR. NELSON: The second is discussion of a bar code label requirement for human drug products and blood. Dr. Lewis?
Bar Code Label Requirement for Human Drug
Products and Blood
DR. LEWIS: Thank you, Dr. Nelson. Good morning.
Today the FDA is publishing a proposed regulation to require bar codes on all drug and biological products. This is part of the Secretary's initiative to address medical errors. As you may know, Secretary Thompson has a long-standing interest in addressing patient safety, particularly in medical errors. The concept of bar coding has been discussed in our agency for quite some time for a number of years to ensure that the right drug is administered to the right patient, and the right dose at the right time.
In reviewing the available technology, the FDA has selected bar coding as a technology that is here and present and that should be implemented if it is possible to save and prevent any medical errors. In the internal discussions of technology, a number of different technologies were discussed and the question whether a particular regulation should be flexible enough to allow different types of technology, for example radiofrequency labeling or other types of available technology. The thought was that bar coding is a simple technology; it is here today; it is in use in many arenas and the implementation of bar coding would encourage its use more widespread.
Also, part of the discussion was if a particular bar code should have a specific symbology that is used. In an open public meeting there was quite a bit of discussion and various viewpoints were expressed on that. The ultimate decision for drug and biological products was that the UCC/EAN symbology, which is relatively common and has some flexibility within it, could be and should be required for product labels.
Who would be required to label products with the UCC/EAN? It would be all drug manufacturers, repackers, re-labelers, private label distributors for all drug and biological products, with the exception of blood and blood components. The particular bar code would be the NDC number for the product. Again, it would be in the UCC/EAN symbology.
For blood and blood components within the FDA we constantly hear the expression that blood is different but, of course, we like to say blood is special and we are special in that we will not require, or we do not propose to require an NDC number for blood and blood components. However, blood is addressed in the regulation and the recognition of the errors that occur in transfusion and the serious injuries that can result.
Studies have shown that the most common error outside of blood banks is that it is administered improperly to patients although labeled correctly. Less common medical errors include mislabeling of samples that are collected for patient typing and cross-matching, as well as the issuance of the wrong unit or testing errors.
A couple of my slides have been dropped or not updated. I apologize. The FDA has published guidance for industry that recognizes the use of uniform labeling for blood and blood components, that is, the ISBT-128 system. In this particular guidance the FDA recognizes this as an acceptable bar coding and labeling practice except where it is inconsistent with the regulations. However, despite international convention and this particular guidance document, there is still not a uniform international bar coding system for blood that is in place in the United States.
In determining the particular component of the proposed regulation for blood and blood components, there was discussion of what type of symbology should be used. As I just mentioned, the UCC/EAN symbology is to be required for drugs and biological products other than blood and blood components. There was consideration of the ISBT-128 and other technologies, and the ultimate decision was that the proposed regulation would require that blood and blood components be labeled with machine-readable information.
Some of the other components of the particular proposal include that all blood establishments would be subject to the requirement that blood and blood components would be labeled with machine-readable information. That would be blood and blood components intended for transfusion and other products from which products for transfusion could be removed.
The information that would be included in the machine-readable format would be a unique facility identifier, a lot number that relates to the donor, a product code and the donor's ABO blood group and Rh type.
The proposed regulation invites comments from industry. I should point out that this is a 90-day comment period so in comparison to many proposed regulations it is brief. We would like your comments on the use of ISBT-128, the use of a UCC/EAN symbology like drugs and biologics, and the ability of particular scanners to read either ISBT-128, UCC/EAN codes or others.
Additional comments are invited on how a particular proposal might affect a hospital's decision to purchase a machine reader. This was a component in selecting for drugs and biologics. It was the reason for selecting the UCC/EAN code. It was thought that by limiting the particular symbology a hospital might be more prone to implement a system that could read one particular symbology and not a vast different array of different types of symbologies. So, we ask that same question and solicit comments from the public and the industry on how that particular part of the regulation might affect the implementation and how the particular types of bar coding might be compatible with other blood labels or other bar codes that appear on drugs, biologics and OTC components.
Thank you very much for the opportunity.
DR. NELSON: Thank you. Any comments? Yes?
DR. DAVIS: Do you envision this being used at the bedside as a point of care confirmation?
DR. LEWIS: No, the intention is that the hospital would use it also to identify patients as well as the blood component. Although the FDA doesn't regulate hospitals and the practice of transfusion, the idea of requiring that these particular types of labels be on blood and other drugs would be that if they were there, a hospital would then implement a system to label the patient and then be able to see, by computer technology, that it is the right patient and the right product.
DR. FITZPATRICK: If the intent is to try to label multi-product and the patient, given the need for look-back information and all the tracking required by GMPs for blood transfusion identification parameters, wouldn't it be possible, without regulating transfusion practice, to require a method of bar coding patient identification? Because most of the facilities are going to say there is no requirement to do this; how do I convince my administration that I need the funding to do it? Yet, if there were a requirement to do it, they might be able to convince the administration to get the funding to do it.
DR. LEWIS: That is an interesting concept. We have been told that we can't go further than we are, but if there were some way to encourage that--we have had discussions with JCAHO on how some of these particular practices might best be implemented.
DR. ALLEN: I applaud your efforts to try to address this issue using systems type technology and I think if we are really going to address the issue of a lot of preventable errors in medicine, we have to do that. Unfortunately, medicine and hospital administration often doesn't think that way and it is, as has been stated a minute ago, a relatively unregulated environment. But I think until we develop a unified systems approach we are not really going to be able to address the issue successfully.
A quick question, do you have any information on whether errors are more likely to occur with the infusion of a blood product to a hospitalized patient, to an inpatient in other words, versus a person receiving ambulatory therapy? That may be an issue that needs to be addressed also in terms of the identification from a systems approach.
DR. LEWIS: I am not aware of studies, although that doesn't mean that they don't exist. The studies that I mentioned this morning are based primarily on some of Jeanne Linden's work where she has shown that of transfusion errors, ABO incompatibility transfusions, about 75 percent of those occur at the bedside and the other 25 percent occur in cross-typing.
DR. STRONG: I applaud your efforts as well and I think this has been the single biggest safety problem in transfusion since the beginning. If you go all the way back to the beginning, clerical errors have always been our biggest problem. But we do have the situation, as Mike points out, that the hospitals are not really ready to spend capital to invest in upgrading information systems to implement this. We have been talking about code 128 now for five to ten years, and I can tell you, having just reviewed the hospitals in our area, that the majority of them still haven't upgraded their laboratory systems to accept that symbology. So, until we have some sort of regulatory pressure to do it, I don't know that it is going to move. There are just too many other financial pressures on hospitals to have them invest in it.
MS. GREGORY: Kay Gregory, from AABB. Mike, in answer to your question about regulatory authority to require bar coding, while FDA probably can't do this, I think the way to do this would be if CMS were to do it and make it part of a hospital's conditions of participation there would be some regulatory key to doing this. Having just read some of the recent changes in CMS regulations, I think they may be moving in this direction because they specifically address upgrading hospital information systems and doing things like this.
DR. FITZPATRICK: Along with your efforts, providing the wrong prescription to the right patient is also a huge problem. So, tying this together with CDER and drugs is a monumental effort and an excellent effort because that error rate is probably higher than the blood error rate and will get more attention, and may get more emphasis for funding if it were tied to that. You might be able to work on both issues at the same time, and it only makes sense to address both issues at the same time and in the same way.
DR. LEWIS: That is one of the reasons that we invite comment on the particular type of symbology for blood and blood components and whether or not machine-readers could read both the symbology required for drugs and biologics as well as those that will be required for blood and blood components.
DR. SCHMIDT: There was a prospective study in Belgium about four years ago. What they found after the blood had left the hospital blood bank were 165 procedural errors for 808 patients, and 15 of them were major. What you had was the hospital transfusion staff who are trained in this, looking for errors, to go out into the rest of the hospital and find out what was going on. It was a prospective study. Unfortunately, I don't think that interest or approach exists in the organ transplantation field where there are highly interested people in the hospital who go out and find out what is going on.
DR. SIEGEL: I just want to comment. I think that people are correctly recognizing that a very broad system-wide approach has to be taken here. There is an umbrella organization, the Agency for Health Research and Quality, HRQ, that is part of the Department and has the mandate to address the whole issue of medical errors for the entire system. There are initiatives in CMS but what we are simply talking about here is the FDA piece. You know, we deal with product manufacturing and labeling and what we are talking about is the product label.
The key message that we are trying to communicate to the public here in soliciting comments on the proposed rule-making is that a specific proposal is being put forward for how to label a blood unit. It is in your interest to think about whether that conforms or doesn't conform to your ideal expectations for where you think you are going with a machine-readable code. So, that is the heart of the message.
Open Public Hearing
DR. NELSON: I guess we are into the open public hearing now. Dr. Dean Elfath, from Baxter Healthcare, wanted to make a presentation.
DR. ELFATH: Mr. Chairman, Madam Secretary, distinguished members of the BPAC, thank you for allowing me to speak this morning.
I am actually here to talk about an issue related to licensing for irradiated red cells because in our effort to design a study that will meet the FDA requirements to license the red cell products collected on the ACDA/ADSOL combination, we found out actually that the original scientific data does not support the current FDA guideline.
The current FDA guideline states that irradiated red cells have a shelf life of 28 days from the date of irradiation, not to exceed the original date of expiration. That means that units that are irradiated on day one or day zero of collection, the day of collection, expire on day 28 of storage. If the units are irradiated on day 14, they actually can be on the shelf for an additional 28 days and that means that they expire on day 42 of collection. If the units are irradiated any day after the 14th day of collection, they still expire on day 42.
So, I summarized the data in the literature regarding the recovery studies, percent recovery of red cells on the date of expiration available in the literature. If you look at the first table, it summarizes the study that was done by Dr. Moroff, Dr. Holme and Dr. Heaton. Dr. Heaton as well as Dr. Moroff are in the audience here this morning. We can see that units were irradiated on day one and recovery studies were done on day 28, the percent recovery, 24-hours recovery after infusion, was 80.2 percent. That is an acceptable recovery because the FDA requires recovery above 75 percent of infused red cells.
If we look at the units that were irradiated on day 14 and recoveries were done on the expiration date according to the current guideline, we will find that they were actually 70.7 percent. That is not acceptable. It does not meet the FDA requirement for 75 percent recovery at the end of shelf life.
You can look at the following tables and basically they show the same type of data. But if you irradiate on day 14 and do recoveries on day 42, you do not meet the FDA standard for 24-hour percent recovery of infused red cells.
So, I am proposing actually that the guideline be modified to limit the shelf life to 28 days of collection regardless of date of irradiation. This proposed standard is actually supported by current existing data, and it is in line with the European standard for irradiation of red cells for adults.
If we would like to add a little more flexibility to accommodate rare types or special clinical situations, I am proposing that the guideline should actually be that irradiated red cells expire on day 28 from the date of collection or within three days of irradiation if irradiation takes place after day 28 of collection. There are no data to support this last proposed guideline but we have an experimental design that will generate the data that we believe will actually support the last guideline.
I think this issue is very important because knowing what we need to meet will affect the design of the experiments that we, as manufacturers, will put in place to meet the FDA guideline. Thank you.
DR. NELSON: Thanks, Dean. Comments? Yes, Harvey?
DR. KLEIN: There are some data, Dean. One of the issues is that not all blood is irradiated in blood centers. Sometimes a hospital, as it is issuing a unit of blood, irradiates it. It goes to the patient's bedside and it comes back. You may want to keep it for another day or two days because of the requirement to use that unit of blood, possibly for that very same patient.
Now, in the last citation you have by Moroff, Holme, AuBuchon and Heaton blood irradiated at day 26 and recovery study at day 28 has a perfectly acceptable recovery. That is only an N of five and I agree that you would like to have more numbers and perhaps more days, but it seems like your initial proposal would restrict the use of something that might do more harm than good.
DR. ELFATH: Actually, that is why I like the second proposal which allows for irradiation after day 28 of collection on the condition that it then expires within three days. I think we can generate data to support that to add more flexibility to the guideline.
DR. NELSON: Any other comments? Judy?
MS. LEW: Could I just ask someone at FDA to clarify if it is true that the data here suggest that going out to 42 days--that there is no data to support going out to 42 days? Where was the original approval? On what data was that gotten?
DR. ELFATH: Actually, I have the original document that includes the guideline, and they acknowledge in that memo that actually there are no data to support recoveries on day 42 but, yet, the guideline ended up being that, yes, you can keep the units for 42 days. But it is not supported by the available data.
DR. KLEIN: Paul Schmidt may remember this, there was a point in time where 70 percent survival was acceptable recovery and 75 percent is really a very arbitrary number. At least 25 percent of the cells are dead within 24 hours of transfusion anyway and 65 percent better or 80 percent better is arbitrary. So, I suspect it may have been at the point of 70 percent that was approved. Paul, you may remember; I don't.
DR. SCHMIDT: I don't either but it seems like Dr. Elfath made a very reasonable suggestion that this committee isn't being asked to do anything about by the FDA. I think he has called it to our attention and we could move on.
DR. NELSON: Jay?
DR. SIEGEL: I think we thank Dr. Elfath for bringing this to our attention and we will review our policy process when we came out with the guidance and consider these data, some of which were known at the time of the decision, I might point out, but, nonetheless, we will reexamine the policy and its basis.
DR. NELSON: Thank you. The next topic is the discussion on the extension of the storage period for pooled platelets. Dr. Vostal with introduce the topic.
Discussion on Extension of the Storage Period
of Pooled Platelets
DR. VOSTAL: Good morning, and thank you very much for this opportunity to present this discussion issue.
What we would like to talk about is extending of the storage period for pooled platelets.
Now, in the current whole blood-derived random donor platelets, this slide outlines the schematic and you start off with whole blood which is then subjected to a soft spin to separate the red cells and the platelet-rich plasma. These are then separated and the platelet-rich plasma is subjected to a hard spin. This brings down the platelets and the plasma is removed. However, a small amount of plasma is left behind to support the storage of these platelets. These are then stored as single units and they are now referred to as random platelet units. They can be stored up to five days. At the end of storage, they are then pooled and they are transfused within four hours of pooling.
The discussion we are going to have today is whether we should pool right before transfusion, as is currently done, or whether it would be better to pool immediately after collection and then store these platelets as a pool. There could be some economical advantages to pooling up front, and that is, if you employ bacterial testing under the current system you end up testing each one of these units. So, you could end up testing five or six times prior to making this pool. However, if we decide to pool up front, it would be possible to bacterial test this pool only once and then assure its safety to the point of transfusion.
So, this seems like a relatively straightforward idea and you may wonder why we haven't actually done this before. This slide shows some of the concerns that we have had that have sort of hindered our progress towards making this change.
First of all, we have always wondered whether there could be a mixed lymphocyte reaction between cells from different donors that could lead to cytokine release and some potential platelet damage.
Similarly, there could be a potential for the presence of anti-platelet antibodies that could lead to platelet damage. There is also a potential for bacterial proliferation in the platelet pool if one of these pooled units is contaminated at collection.
The idea here is that if one of these units is contaminated the bacteria could grow up to its limit to a volume of approximately 50 mL, what the volume of those random donor units is. However, if you combine the contaminated units with five other units, you now have the potential for the bacteria to multiply, proliferate during storage, and you would end up transfusing a lot higher total amount of bacteria to the patient. This concern has actually been demonstrated experimentally so it has always been one of the major reasons why we haven't moved towards the pre-storage pooling.
Finally, we have also had concerns about multiple sterile connections that may lead to a failed connection and product contamination. As you make these pools, you have to use the sterile connector a number of times and if one of those wells failed you could then contaminate the pool which would grow bacteria during the time of storage.
So, why do we think that we are ready to make this change at this time? Well, starting at the bottom, this subject was discussed at the last BPAC and it was decided that the multiple sterile connections are safe and that we can go ahead and proceed at least with this part.
For bacterial testing and bacterial contamination we now have detection devices that are approved, that can be validated for screening these units, and may be able to get around the potential for bacterial contamination and bacterial proliferation in these pools.
For the platelet quality here, we are looking at some of the data that has been collected in Europe. In Europe they have been pooling for a number of years and they don't seem to have any problems. So, one of the things we would like to discuss is whether the data obtained in Europe would be appropriate for the U.S. market.
Our current approach to moving towards extending the storage time for platelet pools would be to have the bag manufacturers validate storage bags for ability to store platelet pools for an extended storage period, and to have the device manufacturers validate bacterial detection devices for ability to detect bacterially contaminated pools.
The process of validating storage bags to store these pools would go through the same process that we usually use to evaluate platelet storage devices, and that would be in vitro studies. This would be testing of pooled platelets at the end of the storage and these tests are outlined in the draft guidance for platelet testing and evaluation of platelet substitutes, released in 1999.
Under normal circumstances we would then move to in vivo studies under IND protocols. These are usually radiolabeled platelet survival and recovery at the end of storage. However, in this situation it may be difficult to do the classical study because there are ethical issues due to random donor platelets given to healthy volunteers and concerns about exposure of these volunteers to multiple donors.
So, we have considered other options and that would be actually transfusing thrombocytopenic patients and following the outcomes of those transfusions. These studies could be designed several ways. One would be using a stable thrombocytopenic population. This population of patients could be transfused with the pooled products and it would be the four-hour pool versus an extended storage pooled product. The outcomes could be a measure of corrected count increments after transfusion.
In these stable thrombocytopenic patients it may be optional to do radiolabeling studies where you actually radiolabel the platelet pool, and since these patients require transfusions anyway you could follow the survival of those radiolabeled platelets.
If there is limited access to these stable thrombocytopenic patients, we have considered going to all clinical thrombocytopenic patients, such as bone marrow transplants and chemotherapy patients. These could again be transfused with pooled products and the outcome would be a measure of the corrected count increments. Study size would probably depend on expected variability to platelet transfusions in this patient population.
As to what are the appropriate controls for these platelet studies, we would think that the right control would be using random donor platelets of similar age, either five days or seven days. The control would be these platelets stored in FDA-cleared storage devices individually and then pooled for four hours prior to transfusion, as opposed to the test group which would be stored as a pool in new devices intended for extended storage of platelet pools. This would mean pooling shortly after collection.
For validation of the bacterial detection devices, this validation would be very similar to what we talked about last time, in December, when we were talking about validating these devices for releasing of platelet products for transfusions.
We would like to see in vitro studies with spiked samples and laboratory studies would define appropriate sampling times for specific bacteria. Then we would like to see field studies with sampling of actual platelet pools. The sampling times are defined by the laboratory studies, and the study design would have a sampling a second time at the release of platelets for transfusion or at their outdate to confirm the results of the first sample used for culture.
The laboratory studies would be designed something like this. You would have a platelet pool that would be spiked with a relatively low amount of bacteria and at some point in storage, say, after day one, a sample would be taken and cultured to produce a result and, at the same time, there would be a determination of the colony forming units/mL to determine how sensitive this device is. The results from these type of laboratory studies could then be applied to the field trial.
The design of this trial would be something like this. It would be sampling of actual platelet pools. You would take a sample somewhere at a point that was identified by the laboratory studies. You would get a culture result and either at the time of transfusion or at the end of storage or even after day six or seven you would take a second sample, culture that result and get a confirmation of the original culture.
Just to briefly summarize what our approach is going to be, we would like to see validation of platelet efficacy, and this would be in vitro and clinical studies. Dr. Ed Snyder is going to talk about these this morning. We would also consider looking at the European experience with platelet pooling, and this data is going to be presented by Dr. Ruby Pietersz this morning. Finally, the issues of validating the bacterial detection devices, the laboratory and the field trials approach will be discussed by Dr. Mark Brecher. Thank you very much.
DR. NELSON: Thank you. Comments? No? Dr. Snyder, we will keep going.
Clinical Performance of Pre-Storage
Pooled Platelet Products
DR. SNYDER: Good morning. Thank you very much for the privilege of allowing me to address the panel.
Our laboratory has done work on platelet storage for many, many years and I would like to present some of the data. Initially when Jaro asked me to talk he said there really wasn't much in the literature but, as always, when you go looking you find a lot. So, I am going to try to summarize a fair amount of material quickly, and some of the slides actually are small because they are scanned in from the original manuscripts but I will walk you through them. It should be visible to the committee but those in the bleachers may have a little trouble. So, it wasn't that I was oblivious; it was a planned move.
Our conflicts of interest--because I am going to be talking about a fair number of commercial aspects, it is important to restate that I am conflicted intensively, which perhaps is the benefit of it--the Pall Corporation, Cerus Corporation, Vitex, Baxter and Terumo. However, I have absolutely no equity; I have no stock in any of these companies whatsoever, although there are paid positions for most of them.
This is "Happy Valley," a slide that I actually love. I have already picked out my house. Unfortunately, we do not live in "Happy Valley" so we need to address a lot of the issues of reality, such as costs of platelet products balanced against the benefit to medical care and so forth. I am going to be living over here, by the way, in case you are wondering.
These are the initial issues which I think the committee should consider for five- to seven-day storage of pooled random donor platelets.
One, is there reason to consider a new standard of care? I think Jaro pointed this out with the bacterial detection devices and that it opens up the opportunity to look at this again.
We need to be mindful of the quality of platelets in the pool bag at the end of storage. We need to be mindful of the effect of multi-donor plasmas on platelets. We need to be concerned about the effect on in vivo recovery and survival because, as we will talk about in a few seconds, all of the in vitro data are suggestive but really the only way to determine if platelets are going to survive is to do a survival study and patient transfusion studies.
Effects on hemostasis also need to be considered, and that is something that has happened with the pathogen reduction trials. There has been a great emphasis on actually looking at hemostatic evidence by looking at patients, which we never really did before we used corrected count increments almost exclusively.
Donor-donor lymphocyte effects to so called admixed lymphocyte culture effects needs to be evaluated. Bacterial contamination issues will be covered by Dr. Brecher. The role of pre-storage universal leukoreduction, should this be built into the concept or not? We will discuss that with some data.
Effect on blood inventory--hospitals are being stretched dramatically. When I gave a list of the upcoming cost increases for a unit of blood that are planned to our administrator last week, I left her in a state of shock, saying that there may be more coming; I do not know. Blood is considered something in the hospitals today that should be used less, and less, and less but you are treading water because the price for each unit and so forth keeps going up, and up, and up because of all of the changes. So, I think inventory effects and the increase in wastage is not a trivial issue and it does relate directly to the public health, the ability to have a product there when a patient is actively bleeding.
I come from Yale-New Haven Hospital. That is all I have been, in academic blood centers, and this is what we have done over the years. This is till 2002. We transfuse about 43,000 blood products a year.
To approximately 7,200 patients a year, 7,300.
We have only used random donor platelets. So, this is an important issue for us. We do not use very many single donors, we never have. When Joseph Bovy was the director we had 12-unit pools and then we went to eight, six, five. Due to a variety of budget reductions, and evidence-based medicine that really were needed, we are currently at a four-unit pool. Next year we are planning to go to a per platelet basis, not units of platelets but individual thrombocytes where we would actually charge--
In the future we are considering the virtual platelet transfusion. We would just show them a picture of platelets and charge for my time.
This is facilitated by a product that is under evaluation--
--the "platelet helper." Just add platelets for faster clotting. This is my favorite flavor but it does come in a variety of other flavors. That is the last humorous slide.
Random donor platelets--we have done our best to decrease the number and we use about 12,000 units of random donor platelets. We use 40 unit pools so there are about 4,000 to 5,000 pools a year. We waste about 10 percent, which is close to 1,000 units a year, because we have to have platelets on the shelf. We are the so-called 800 lb. gorilla. We are expected to have platelets to give to other hospitals when there is need or shortage. We also have to have platelets for the helicopters that land on the roof, etc. For us to lose about 10 percent or more of our platelets because we pool for the operating room and then it comes back from the OR unused because when things happen in the operating room everyone starts screaming--you try your best to make sure that they don't over-order but it happens and it comes back. If we sent it out as a closed system pool and it came back as that pool we could put it on the shelf and reuse it. This happens sometimes more, sometimes less. So, there would be a very practical reason why a closed system pool that was safe would be of benefit to our institution.
This is single donor. This is in hundreds. This spike actually was due to a study we did for a corporation but we are back to about 100. Most of these are atrially matched products for specialized patients.
What about the platelet storage lesion? Well, permit me to give a little bit of background for some of the members of the committee that may not be as up to speed on this so that when you see the slides it will be a little easier to understand.
The storage lesion refers to any untoward effects on platelet structure and function during the collection process and storage as well. The mechanism of action of this is multifactorial. Every organization, group and committee has their "Holy Grail." Our "Holy Grail" is to try to find an in vitro assay that will tell you what is going to happen in vivo. That apparently just doesn't exist because it is so multifactorial that we are really not quite sure what is causing it.
The bad news is that in vitro assays, per se, are not very predictive. The good news is that when used in conjunction with radiolabeling and in vivo post-transfusion assessments you can get good data. If all the in vitro data you look at by the criteria and the radiolabeled survivals are good and the transfusion studies are good, then you know you have a product that you can license with confidence.
The guidance for industry that Dr. Vostal talked about came out in May, '99 and I understand an update is in process. But there were four categories of evaluation. One was in vitro using three categories, morphology, biochemistry and function of platelets, and you will see some of these in the data.
In vivo platelet survival in the circulation includes radiolabeled survival with indium or chromium, as well as post-transfusion corrected count increments.
Clinical hemostatic efficacy is an issue, which was very important for pathogen reduction, that was mentioned as being important in this document and it was raised for consideration as to whether it is needed for this study, and I will give you my ideas on that.
Then, evaluating platelet substitutes really related more to other products and is not germane to this discussion.
The assays that are being used--this is taken from Scott Murphy's Transfusion Medicine Reviews. There is the metabolic group, pH, which is probably the one that everyone looks at the most with the concept of swirl which is related to when the pH drops or the platelet undergoes a shape change from a disc to a sphere, and when it is in a disc you get refraction so it looks like it is opalescent; and when it undergoes a shape change it rounds up and you lose this opalescence and it loses this so-called swirl. So, that is one issue that is looked at.
Then, you should see changes in pO2 which fluctuate widely. You can have a very wide range in pO2 and still have a reasonably viable platelet and pCO2 and bicarbonate are linked together for obvious biochemical reasons. The platelet count is important. As you metabolize through the anaerobic pathway as opposed to the creb cycle you build up lactate. You do not want a lot of lactate because that lowers the pH. If you go through the creb cycle you have CO2 which can be blown off because of the fact that the bag is porous and glucose will be consumed as well. So, you want to see these numbers to be fairly low. This number low, rather, and this number high as compared to day one.
Morphology, generally you don't really need to do fancy electron microscopy work. Then there is LDH, which is in the platelet cytosol. If you see a lot of LDH release it means the platelet is breaking up. So, you want to see a low number here. Beta-thromboglobulin is in a granule in the platelet. You want to see low levels of B-TG if you measure irradiated.
CD62P is actually a marker on the granule membrane surface. When the platelet undergoes release that gets to the exterior of the membrane and it can be detected and seen under flow cytometry. So, if this number goes up, this number should be up as well and we have general guidelines for where they should be, somewhere in the 20 to 30 to 40 percent, hopefully.
Hypertonic shock response is the ability of a platelet to pump out water when it is exposed to water. Extent of shape change has been studied extensively by Holme and Murphy, Moroff and others. The feeling is that the shape change that occurs in response to ABP may be more beneficial in predicting what happens in vivo. Then there are some other values, platelet volume, some other factor assays which are really not very often used.
Then there are more esoteric ones which are not used, EM morphology and aggregation, which is beloved by the FDA. We refer to that as platelet aggregation studies. That was from Ted Spade actually, in New York, who coined it for me. There are a variety of other glycoproteins, microparticles, things which are not usually done.
Several of these slides come from Kathryn Webert, who was a Fellow at McMaster, who presented a talk on platelet pooling at the AABB and graciously gave me permission to borrow some of her slides for this talk, and I have indicated them with her name at the top of this.
This just shows the difference between the platelet-rich plasma method which is used in the U.S., versus a buffy coat method, and Dr. Vostal discussed that previously so I will skip that in the interest of time.
First in vitro evaluations comparing pooled platelet random donor concentrates with unpooled.
This is a summary that Kathryn had. There were several studies. I am not going to go into great detail in all of them; I picked several of them, Bertolini, Moroff, Wagner and Heddle. Bertolini's paper looked at unpooled units versus pooled. She stored platelets up to 15 days. This just shows you the different types of protocols that have been used. They stored them in PBC bags and did not leukoreduce.
Moroff's paper in '93, pooled versus unpooled, five-day storage in 732 bags; was leukoreduced. Steve Wagner's paper, pooled versus unpooled up to five days; did not leukoreduce. Heddle, with an abstract actually, using a different type of bag did leukoreduce. So, it is kind of all over the map. There is no large amount of literature using one standardized technique.
This is the top of Dr. Moroff's paper from 1993. I just wanted to bring this because I was practicing scanning things and doing very well actually. The bottom line is these studies demonstrate that pools of ABO identical PCs have satisfactory in vitro platelet properties after storage.
Let's see why he said that. What Dr. Moroff's group did was to take pools of platelets and store them in six-unit pools for five days and compared them with individual units that were stored for five days. Then he looked at eight-unit pools as well and compared them with individual units that were stored for five days as well.
Here are the various variables. Let me just walk you through this briefly. The platelet counts here are six times what these are; these are individual units. Six times 0.83 is roughly 4.7, and the same with the eight-unit pools.
These were not leukoreduced, as you can see. This is 1 X 109. This is a single unit over here so this is six times the amount. pH is important. pH at five days for the pool was 7.12; for the individual units was 7.14, 7.09, 7.2. This means that at least crudely pH is not falling when you store pools of platelets together even without leukoreducing.
Morphology scores, out of 400, 290 versus 292 and 263 versus 263. Ideally it should be compared with day zero and some studies have done that but it wasn't done in this particular study.
Hypertonic shock response, 50 percent recovery is very nice; 52, 49, 50, 44. The standard deviations here are fairly wide because of biologic variability so this is perfectly acceptable clinically.
Beta-thromboglobulin release, 26 percent, 27, percent, 23 percent and 30 percent, no problems there. LDH, this is roughly a number that is compatible with a very intact platelet that is not undergoing lysis and so forth. So, this seemed to show that there were really no problems here.
He also pooled platelets that were prepared on day three instead of day one. This showed the platelet count as you see here. pH was 7.24. These were stored for an additional two days just to see if it had a problem if it was pooled later, and it did not with morphology. Shock was fine; volume was lovely. So, there was no problem here.
This is the second laboratory that did essentially the same thing. Just let me point out a few other things. There was really nothing different. The six-unit pools and eight-unit pools versus single units stored for five days showed no significant differences.
For lactate here you can see 15, 14, 14. So, that implies that the pooled units versus the single units were not generating more lactate because there was a problem. Glucose, 14, 18, 16 with appropriate variability there. ATP was 7.22 versus 7.28, which was lovely. The CO2 showed really nothing out of the ordinary. So, his conclusion was that there was no problem and I have just shown you the data.
Nancy looked at P-selectin. A lot of people didn't look at P-selectin because when these studies were done we didn't have P-selectin. It was called Paguin actually, but I digress. This is in Nancy's abstract. She looked at pooled versus unpooled platelet concentrates for generation of P-selectin--I am sorry, CD-62, the same thing, and it should correlate with beta-thromboglobulin because the CD-62 and the P-selectin are the same protein on the membrane of the alpha granule; BTG is what is inside the granule. A 40 percent level of P-selectin is really not a problem at day five. Again, you would like to see a nice standard of stored/unstored. It was in an abstract so I don't have any more data than this.
Bertolini looked at collagen. Presumably he looked at collagen 50 mcgm and 10 mcgm. When platelets are stored, at least with the PRP method, they really don't like to talk to agonists. They are really not very happy; they get less happy as they get stored so you have to go to dual agonists ofttimes. This was just showing that if you stimulate platelets with collagen 50 mcg/mL you get about 54 percent aggregation with the pooled, 55 percent unpooled on the first day of storage and if you used a smaller amount of collagen you got smaller levels.
Most of what Dr. Bertolini did was comparing his work in a glucose-free media with buffy coat versus platelet-rich plasma. Again, there wasn't a lot of data but I thought this was useful because it had collagen in it.
For pH Bertolini again, pooled 6.7 and unpooled. The acceptable range down here, from what Kathryn wrote, is 6.2 to 7.6 for platelets. I kind of disagree with that. Although the 6.2 is the lower level, we like to see something that is probably not much below 6.7 and certainly not much above 7.2. I think 7.6 is way too high because you start getting the release reaction occurring if the pH gets too alkaline. But here Bertolini's was 6.7, 7.1; Moroff, 6.96 and 7.10; and Heddles's group was 6.8 and 7.0. So, there is really no evidence so far that there is any problem with pooling with a variety of different techniques.
In summary of this group, the morphologies were found to be higher in one study. If you do enough evaluations of enough assays you are going to find something that is significant. But they were all within acceptable ranges so you get into the not clinically significant but statistically--osmotic reversal was not significantly different. The parameters reflecting platelet activation and lysis were not different. in vitro metabolic parameters were not different, and the aggregations studies were no different.
So, from the evidence from in vitro assays here it would appear that pooling platelets is not a problem, which is good news.
What other studies are there? Well, evaluation of pooled platelet concentrates using pre-storage versus post-storage white cell leukoreduction: the impact of filtration timing. I chose these papers because these were papers that looked at pooled products in addition to doing other things with them, which included filtration.
This is one of those challenging slides for the folks in the back. This was a group that pooled large numbers of random donor platelets together. Then they either filtered them immediately or they let them store for five days and then filtered them later. So, there was an early filtration pool and a late filtration pool but it was all the same material. They pooled and split the product and filtered one group before and after filtration on day one, and then on day five they took the other half and they filtered it and stored it for up to nine days.
What we see here, we are comparing pre- and post-filtration, which basically looks at the effect of the blood filter which is not what we are really interested in. Looking here at day five for the early filtration pools, if we look at important things, pH 7.1 so there was really no difference in pH after five days of storage before filtration or day five of this pool of platelets which is I think about six or seven units.
pCO2 showed a little bit of a drop, which was statistically significant from day one but not clinically of importance. The bicarbonate shows it started to drop but that is a function of storage time; not a function of pooling in my opinion, so I don't think it shows any difficult problem. This was pooled with a PXL-8 pool filter. Glucose started to be consumed a little bit, as we see during storage; no big change there.
There was an increase in LDH but, again, 130 plus/minus 131 compared to 299; no change there. If you look before and after with the late filtration--and I don't recommend that these products be filtered very late; I think they should be filtered up front for reasons we will talk about in a little bit--this is the early filtration pool and the late filtration pool and what they found was pH 6.9, 6.9. So, even out to nine days the pHs were maintained in large pools of random donor platelets.
The CO2s were not excessively high. If the platelet had died the CO2 would actually go to zero because it would diffuse out of the product and the oxygen would go way up to like 190, 200 because the bag would become basically just a place for air to move back and forth; there would be no metabolism going on. So, the fact that the CO2 is in a similar range, even though it is a little lower, no big problem.
There was consumption of bicarbonate, again a storage-related problem. Glucose was consumed compared to day one but no big problem there because day five certainly looked fine, as did the bicarbonate on day five. Then, LDH went up a little bit, as you might expect for a platelet that is having its little membranes beaten up 70 times a minute for nine days. You would be beaten up and have your LDH up too if it happened to you. So, this didn't show any problems.
They looked at other things and they looked at CD62 before and after filtration of the pool. This is again the percent of the alpha granule that is on the membrane and 19 percent is lovely. Looking at day nine, 39 percent does not show to me any excessive platelet activation.
They looked at CD63 as well, granuloficin, and that went up as well a little bit but, again, nothing severe. Then they looked at GPIb and GPIIb3a with monoclonal antibodies and there was no change in that. So, the platelets are showing the same thing over and over again.
They also looked at thrombin generation with prothrombin fragment 1 and 2, and thrombin/anti-thrombin complexes, and there were no changes really during storage showing that thrombin wasn't being generated with pools.
So, you are looking at all these different kinds of things and it is all telling you the same kind of thing, you know, not to worry until we hear from Mark and then maybe we will have to worry.
We can skip this one and the next one. Those are about the individual filters.
There was a paper from Dr. Boomgaard, from the Netherlands, pooled platelet concentrates and filtered with three different filters and stored for eight days.
This relates to if we want to filter these, is there any one filter that might be better than another. They looked at a Pall, a Sepacell and a Bio P10 filter. There were appropriate amounts of leukoreduction. They had a massive amount of platelets pooled together and divided them into four. Three of them were filtered and one of them was not filtered. So, everything was pooled. There was no unpooled compared to the pools.
Here they stored them for eight days. Here we look at pH. If you go across, all of the filters started off about 7.1 and they all ended about 6.9. So, there was no problem with pH.
CO2s were all lovely, 59 going down to 49. pO2s in the 100s, very nice. Swirling--this was after Fratantoni's scale. I am not familiar with that. I would assume it went up to 4 being the best because it dropped down to 1.8. It happened consistently and I am sure it was a storage-related characteristic but there would presumably be some less swirl but still some visible. Morphology scores, out of 400 there were really the standard drops. This is really what we see.
Filtration with a whole bunch of filters, some of which have positive charges and some of which have negative charges on the membrane as the mechanism if white cell removal didn't appear to affect the platelets.
Looking at glucose levels and lactate generation, they did it sort of in a 1:8 nanomole consumption per day and there was really no difference among them compared to the unfiltered group. Adding up all the AM, AD and ATPs there were no differences there except for some drops as you would see in storage. The beta-thromboglobulin levels seemed to be fine over eight days.
They also did aggregation studies. This was with ADP and collagen, and all these really show--this is light transmission so it is the amount of aggregation--over time this is what you see but there was no difference among the various filter groups. The dotted line is the unfiltered control. So again everything seemed similar; no problems.
We can skip this, and we can skip the next one as well.
I asked all the companies if they had any data that they wanted me to present. I got this from Cerus Corporation. I show it because they are pooled studies. This was published in Infusion Therapy and Transfusion Medicine, which I understand is a German journal. They took two units of single donor platelets and divided them into two. So, they pooled single donor platelets, not random donors, and they were treating them or not treating them with their S-59 device. I mention this only because the field is now also looking at pathogen reduction. The question is, if we adopt this, is there any data that pathogen reduction technologies such as are available would have a negative impact on the pooling process? What data is out there that is discussable, and this is why I am showing this.
The control units did not receive S-59 and were not illuminated, which is part of the S-59 process, and were not incubated with a compound absorptive device which removes any photo product that is remaining. They were stored at 22 degrees and samples were taken on days two, five and seven.
So, they are looking at pooled control or treated with the S-59. The summary of all of these various morphologies, etc. was that there was really no difference. This looks like the error bars are too high. So, it would appear that pooling at least two units of single donor product with or without S-59 treatment up to seven days did not show any differences in vitro analysis.
They also did a study coded as DELL-240, which is pooled random donor platelets. Here they took individual random donor units and they pooled them. Seven units were pooled and filtered, and they took out the red cells.
The random donor platelets met the criteria. They had 2 X 1011. The S-59 is added to inactivate viruses, bacteria and things that have DNA or RNA. It will inactivated it once ultraviolet light is shone on it. So, the controls were made by pooling five conventional concentrations and giving them four hours after pooling. In addition, an important point is that the pools of the product that were in the test were stored PAS III. PAS-III does not have glucose. It is a material that is used to store the platelets that has bicarbonate, phosphate acetate.
After they added the S-59 treatment it was illuminated and then they were stored and samples were taken on day five and seven of storage.
What we have is day five control, which are units of platelets that were pooled within four hours; they were evaluated on day five. Then we had treated groups that were pooled at the beginning and stored for five days, and also sampled on day seven. So, these were stored for five days, individual units; these were pooled and stored as a pool for five days, and a pool for seven days but treated with the S-59 technology.
The pHs, over here, essentially are statistically significantly different. I did not get from the corporation any p values so I will just give you my impression that these are lower values, and certainly day seven is somewhat lower, but they are above the 6.2, 6.3, 6.4 area. Although this is getting a little low for my taste, it is still acceptable.
CO2s, 51 is fine; 33 is acceptable on day five; 26 is also acceptable. There is a wide range in pCO2 and 122 is fine; 61 is fine; 46 is a little on the low side but 61 a day or so later. Bicarbonates were low. Among the reasons the bicarbonates were low is there is no bicarbonate in PAS-III, which is the material the platelets were stored in, so they were low already.
Platelet counts are as you see here. Morphology scores were similar. Extent of shape change was statistically significantly lower, I am sure although I don't have a lot of data. This seemed similar to the control. Hypotonic shock response, there seemed to be no problems across the board. ATP seems a little lower than on day seven.
P-selectin values were higher, which could be due to the fact that there was extra manipulation going on during the generation of this treated product, but you would have to do sham studies to validate that. Lactate levels here are 13 and 17, showing similar levels to the 16 in the control. Glucose is much lower because there is no glucose in PAS-III. LDH seemed somewhat higher but not excessive amounts of hemolysis.
The conclusions I got from this are that for this particular study I didn't see anything that I would interpret with the S-59 treatment of a pool as showing that this was going to be a major problem, but obviously a lot more data are needed and this wasn't analyzed statistically.
What about MLCs? Well, let's go back and look at some data. This is B. Margo and Sonny Dzik. This is mitogen response of lymphocytes and platelets. What Sonny and Margo did was look at--I think the next slide will show this a little better.
They took platelets that were either fresh or stored right after preparation, random donor platelets, or stored for three days and treated them with three mitogens in products that were collected in CPD or CPDA-2. CPDA-2 is one of the four versions of CPDA that were originally evaluated but two, three and four were never licensed, to my knowledge, and just one made it through; two, three and four went bye-bye. But at the time apparently it was available to them.
What they did was tritiated thymidine uptakes, which I think are familiar to most of the members of the panel. The concern is that the lymphocytes in the blood bag, and this was a non-leukoreduced product, would be stimulated by a mitogen and would undergo blast transformation. If it did, it would incorporate DNA materials or nucleotides from the environment. If they are radiolabeled and incorporated, you would then see how many counts there are. What he found was that there are 99,000 counts on day zero with CPD; a similar amount with CPDA, showing that adenine really didn't seem to have much of an adverse event, but by day three the count numbers were a lot lower. So, he was looking at the different mitogens. Here is 50,000 versus 41,000; 31,000 versus 37,000. There were no differences between the various storage preservatives or the various mitogens. Comparing day zero to day three, there were statistically significant drops, which they concluded was because of the storage. That is, as white cells store in a blood product they actually become less responsive, less immunologically reactive, which is a good thing. We don't want them going in and blast transforming and doing nasty things. This was greater than 0.05 just because the numbers were small. It was an N of 10 and it just didn't quite reach it, but it went from 50,000 to 21,000.
So, this was sort of the first shot, as Sonny often does. He often gets there first and has very important data to come up with to get us thinking, and he succeeded again this time, showing that there really didn't seem to be a lot of blast transformation concern with this type of a product but we don't put this stuff in blood bags.
What other data are there? Drs. Vliet, Dock and Davey looked at factors in the liquid portion of stored blood that inhibit the proliferative response in mixed lymphocyte cultures. They took units of blood and stored them for up to 14 days, and took the supernatants from either CPD or CPDA--CPDs are ADSOL, and took the supernatants up to 14 days, sampled them at different times, and added them to MLCs from normal donors, looking to see if there was something in the supernatant that would inhibit an MLC reaction. Was there something being generated to stimulate an MLC reaction?
I am going to skip this and come back to it. I want to put that towards the end.
The first thing they did was to take a look at 14 days of storage, with any changes in the mononuclear cells in any of the antigen typings for CD8, CD4, CD15, and they basically found that there was no difference during storage. They were not looking at changes in lymphocytes during storage. They were the same.
Then what they did, again, they took normal donors. They irradiated one of them so it was a one-way MLC. They had eight responders and they took samples up to 14 days and added them to the MLCs. What you see here is the percentage of relative response which they state is the net counts per minute in the test sample divided by the net counts per minute in the control. So, if there was no inhibition of the supernatant taken from these stored blood products, you should see 100 percent here. In black is CPDA-1 and in the grey is ADSOL.
As you can see, in all of these responders there appeared to be a greater inhibition due to the CPDA than there was to ADSOL, but in none of them was there 100, except one or two. Most of them were fairly low.
This slide shows when they grouped this together by day of storage. Now they took all eight responders and they added all their values together. They did it for day zero, day 7, day 14, and day 28 of storage. They showed that whatever this inhibition was, it occurred right away at day zero, which meant that you didn't really have a lot of time to generate cytokines and things like that. So, they are postulating that maybe it was plasticizer from the bag. Some people said, well, maybe it is less calcium because you need calcium. The material that was added from the stored blood supernatants was recalcified and heated. So, it wasn't complement. There was calcium there. They didn't have an answer. They don't know what it was that was inhibiting it.
They took MLCs again. They took five of the responders and they added either 10 percent normal human serum and they got counts of 18,000, 30,000, 38,000, 53,000, 44,000, showing some variability. They added adenine at 0.27 mg/mL which is the amount that would be seen in an ADSOL unit and they did find some inhibition when they added adenine. So, now they are taking normal human serum and adenine and adding it to a one-way MLC. They are not using the supernatant of the stored blood anymore; they are actually adding some products.
They wanted to check whether these were responsible for the fact that there was inhibition. They found that adenine did inhibit it somewhat, 24,000 versus 44,000, 37,000, 53,000, but they didn't find it with a lower amount of adenine, 0.05, or with varying amounts of dextrose, 22 mg or 6.3 mg. There were no differences here. All they found was that high dose adenine seems to have an inhibitory effect on platelet MLC but no one has really found much in the way of MLC.
Going back to Dr. Moroff's paper, he also did MLCs. I didn't mention it before because I was holding back the good news. He did tritiated thymidine uptake from five-day stored pooled an unpooled platelet concentrates. Here are the number of units. This is the thymidine uptake. There is really nothing there, 267 counts for the pool. For the individual units, basically 160, 250. Then a positive control was done which had the 10,000, 20,000, 50,000 that you see.
So, clearly, lots of studies from different groups--and there is more to come--haven't shown that there really is an MLC occurring in the bag, for whatever reason. We don't really know why it is but it doesn't appear to be a lack of calcium. It may be something that is due to the plasticizer. Now I think we also know that it could be due to storage because when you store you lose some of the cofactors that are required, B7 and so forth, that are necessary for antigen stimulation.
Dr. Webert concluded that comparable low levels of thymidine uptake were detected in the mononuclear leukocyte fraction of pooled and unpooled stored for five days, indicating that mixing lymphocytes in the pool did not stimulate in vitro immunologic reactions. Again, these are not leukoreduced products so leukoreducing would add another layer of comfort. That is why I gave you my conflict of interest statement, I do believe that is true but you need to know about the conflict aspects.
We will skip this because this relates to the work we did, which I will get to next.
Now in vivo evaluations comparing pooled random donor PCs with unpooled--
Here I am. Apparently the only other paper that may be in the literature, and I am not sure, is an abstract that Gail Rock published in Transfusion as an abstract where she may have pooled random donor platelets as well. I am not sure. I couldn't get hold of her before the meeting and I am not sure from reading the abstract exactly if she pooled and stored them as a pool. I don't know. They may have been stored and then pooled later on. That is in the bibliography that I sent you, her abstract in Transfusion last year.
We did a study, and this was in 1989. We took random donor platelets. We pooled them in four- and five-unit pools, both ABO compatible and ABO incompatible. We stored them for up to five days and looked for changes in in vitro storage. We stored them in PL-732 bags.
We can skip that.
This is zero hours, four hours, 24 and five days. C is compatible; I is incompatible. We pooled As and Bs together or just Os together. These were non-leukoreduced products. The white counts here are 1 X 106/mL. LDH, going from zero hours, 7.4 percent. So, we took the LDH in the supernatant of the platelets, spun supernatant, and we took the total LDH by Triton lysing the platelet-rich plasma to get the denominator and we expressed that as a percent so we didn't have to worry about platelet count. So, 7.4 percent versus 8.0 with the compatible versus incompatible pools at zero hours; 7.4, 8.1 at four hours; 24 hours--there was really no lysis going on.
Then we looked at beta-thromboglobulin release, 18, 19, 18, 20, 20, 21, 27, 25, nothing except a little minor trend due to five days of storage.
pH, 7.3, 7.3, 7.3, 7.2. pH started to fall and these units had about 3 X 1011 in the bag so we were pushing what a PL-732 bag should be. When we did the in vivo transfusion study we changed to a one liter bag to get better oxygen gas exchange.
CO2s, 26, 26, 29 and 44, no problems; the platelets were viable. O2s were a little low here at four hours, and then they stayed somewhat on the low side but that was probably because we had them in a 732 bag and it was close to bulging.
Lactates showed some elevation by five days, again showing it needed a little more oxygen. Glucose was consumed as well over five days, but no difference between compatible versus incompatible. Osmotic recoveries went down a little but were in a good range.
We also looked at C3a and saw no generation of C3a during storage. Complement activation wasn't occurring. We don't really measure C5a. We are actually measuring SC5b9 which is the footprint because the body doesn't like C5a floating around; it does mean things so it tends to get bound quickly. It was under the low control of the test kit that was used at the time.
We then looked at MLCs. We did a paired study where we took three donors and made random donor platelets from them, and then took those and pooled them. We then took the same three donors and took their platelets separately that were not made into platelet concentrates but just taken out of the arm, and we used that in an MLC. The MLR refers to the three donors whose lymphocytes were taken, were stored at 37 degrees with NHS and treated, looking for generation of T-cell activation markers, versus the same three donors in a pooled platelet concentrate MLC, trying to say you are not getting generation of MLC but could it have generated MLC at all? I mean, are those lymphocytes able to do anything? So, this was a control for that. This is zero, 72 and 120 hours worth of storage. There is no error bar because it was an N of 1. That is the problem with the study; the numbers were too small but we wanted to get it out and it needs to be repeated.
What we found was that the pooled concentrates, again, did not generate IL-2 receptor, transferring receptor or HLAA DR on LEU3 positive cells. There was really no fluorescence. But the MLR from the same donors that were done under conditions that were not platelet storage, you were able to see a generation.
We also did tritiated thymidine uptake under the same conditions. Here are the pooled which showed no uptake, as others have shown multiple times. We found that by day 72 you had a very strong MLR, clearly showing that the lymphocytes were capable but something about preparing a blood product inhibited that.
We also looked at other types of platelet conditions, a four-unit pool that was stored up to seven to ten days, individual units of random donor outdated four-unit pools. We couldn't find any situation where there was a generation of a T-cell marker in those stored products, just looking at them at random.
Lastly, the in vivo study where we took platelet pools, pooled them at day one; stored them for up to 96 hours. Prior to that we went to the IRB and got approval of the IRB for informed consent, written consent. Five patients were transfused in the study. They were actually given platelets. Gram stains were done, both gram stain and acridine orange plus a visual inspection. Corrected count increments were done one to two hours and the pooled were 11,000. This was a crossover study so these were the same people. This was the so-called standard sick thrombocytopenic patient. They were patients on chemotherapy who were not actively on chemo. They were not bleeding. They did not have big spleens, etc., sort of the standard condition. At 12-24 hours the CCI again was better for the pooled stored. There were no transfusion reactions associated with this, thank heavens, and everybody did well.
The overall summary was that storing as a pool is at least equivalent to individual units from all of the data, again summarized by Dr. Webert. There is no evidence of lymphocyte activity. There is some evidence that pooling may result in greater numbers. She was referring to Dr. Wagner's and other's work which will be presented later. There was no difference between pooled and unpooled.
Issues to consider are the issues that I mentioned before. Let me give you my thoughts on them very briefly so you can take a break.
The quality of platelets in the pool bag at the end of storage is adequate based on published data. I believe that is correct so I think that that is in our favor.
More in vivo studies are needed, as Dr. Vostal points out, but I think the data out there shows that it is not a fool's errand. There is good evidence to believe that the in vivo data will show that this is doable if we can get the main concern, which is bacterial contamination, under control.
The effects of multi-donor plasma on platelets is of little concern. Remember, I used compatible and incompatible pools. We did find that there was clumping of little bits of a few red cells that were in the pools. We actually did heat elutions on them and eluted off anti-A or anti-AB. There was less than 1, 2 drop in antibody titer. I don't think it is a big problem. I am not saying you should pool and store ABO incompatible. I don't think it should be done, and we didn't transfuse ABO incompatible pools but I don't think there are any big problems but I wouldn't recommend it anyway.
The donor lymphocyte effects are not a problem. There apparently is no bag MLC. Buffy coat pooling in Europe I believe is supportive of the process and I don't feel we need to have extensive hemostatic evaluations which require hiring large numbers of research nurses to go to the floor and extend these studies out a year or two. That is my opinion.
Leukoreduction is beneficial. I realize there is not a lot of MLC but I would recommend if we are going to store pools leukoreduction should be used. That is my personal opinion, without any of the conflicts, because I think it just gives you an extra level of safety for storing a pool.
In vitro patient transfusion data are needed. Radiolabeled normal volunteer donor studies are unethical. You can't transfuse a pool into an individual. I tried to think of a good way you could do it, maybe if they are identical twins but then it is the same platelets. I gave up. You just need to do patient transfusions.
Bacterial issues will be addressed. Some pathogen reduction technologies appear compatible; others need to be evaluated. A partial extension is better than none. If you give us 72 hours of storage, that is okay, but don't give us eight hours because an eight-hour extension doesn't do anything because we can't get the bacterial testing back and eight hours is nothing. So, if the concern is that five or seven days is more than you want to do now, three days would be all right as well from my perspective, just from my perspective, and the public health needs would be served because we could reuse those pools and we wouldn't waste it and there would be more for the patient.
I will stop there and answer any questions if you have them.
DR. NELSON: That was a rather large amount of data to digest, but interesting. Yes?
DR. SCHMIDT: The question of room temperature storage I think has been taken care of--you know, what is room temperature. It was a mess at the beginning. We found out that cold is bad and I think room temperature has finally been decided at 20 plus or minus 2.
Just going through some of the material presented to us while you were talking, sometimes they don't say what the room temperature is. We had to buy little pieces of apparatus which made a room inside a room but, anyway, maybe our next speaker would tell us for certain that room temperature in Sweden, where a lot of these studies were done, relates to this because I presume it has been stabilized but it was a mess at one time.
DR. SNYDER: The studies that were done at our institution, certainly because we were AABB accredited at the time they were done, we monitored the temperature of storage and they were 20-24 degrees. Most of the studies actually did state the temperatures and nothing popped out that they stored it in the cold or 17 degrees. I think everybody did the right temperature.
DR. SCHMIDT: The document from the FDA that we got says U.S. holds at room temperature. It doesn't give the temperature.
DR. SNYDER: Right.
DR. GOLDING: Just a few immunological comments. I think your own data actually solves the mystery of why you don't get an MLC in the bag which was stored at room temperature, and the solution to that mystery is probably the temperature. When you do the same cultures at 37 degrees you get very nice MLC reactions.
The other point I would make related to that is that if you add calcium in the presence of the right temperature you would also get a good mixed lymphocyte reaction.
But apart from that, I think there is a concern here and the concern is, sure, the leukoreduction removes intact white cells, not all of them and you still have some white cells. My question is how many white cell membranes remain in the prep. and what is getting into the patient, not what is happening in the bag, and in the case of pooled samples could that still be a problem in terms of getting an allogeneic response?
DR. SNYDER: That is a valid point. It is one of the reasons why I personally prefer pre-storage because you have less microparticle formation by getting it out early rather than filtering it later. You still have some; I don't deny that. Your points are well taken. I don't have good answers, but I think studies need to be done to take a look at that.
DR. KLEIN: I may have missed something here. Ordinarily you are pooling them in the recipient so really what we are worrying about is what is happening during the course of storage. I didn't see any data or hear any data to suggest that there is something nefarious happening in the period of storage regarding MLCs, unless I missed something here. So, I am not quite sure what the issue is in terms of what is going into the recipient. All those white cells go into the recipient anyway whether you pool them at the beginning of storage or whether you pool them at the end of storage.
DR. SNYDER: You are right. It is a bit of a complex issue. I guess the concept would be that if there is an inhibition in vitro maybe you are releasing it during transfusion and you are getting a donor-donor reaction inside the patient as a vessel.
The other issue to bring to bear on this is one that many of the products given to patients with oncologic disorders are irradiated so that you are inhibiting it from there, whatever is left.
The second thing is, and maybe Dr. Pietersz will speak to the issue because they have a lot of pooled buffy coats that have been used in Europe, tons, and I haven't seen tons of abstracts coming out about MLC reactions sweeping the European continent. So, hopefully, there is some information on buffy coats that will be brought to bear on this.
DR. LEW: I noticed that in some of the studies there was a lot of variation of what was considered statistically significant. I don't know if it is just because the numbers are so large, but it varied from a p of 0.05, the standard, to 0.01. If you could just clarify that because I am wondering if it was always 0.05 and everything looked statistically significant so you just had to back off.
DR. SNYDER: It depended on the investigator. There really was no standard approach. Nancy Heddle used 0.1, others used it as well. Your point is well taken. There is not a lot of standardization so I can't comment any further on that, but you are correct.
DR. FITZPATRICK: Dr. Snyder, thanks. That is a great review of the parameters. I was hopeful that I would see more about the effects of hemostasis because what we saw was essentially survival and circulation and, to me, that is not really hemostasis. That is survival and circulation. That has been a problem with platelets forever, how do we show that they are effective when they circulate? We really don't have a good measure of that that is repeatable, doable. I would be happy to hear it if you can come up with something that would be more beneficial.
DR. SNYDER: I don't. Your point is extremely well taken. Somebody may say the word "bleeding time" but it wouldn't be me. I think the feeling of the committee and the field is that we need to do hemostatic efficacy and then you need to really trot out the World Health Organization evaluation and do a full-scale evaluation of this. Hopefully, we could generate enough data with buffy coat from Europe that we may not need to do this, although it may be apples and apple sauce; it may not be exactly the same but similar. I would leave this to the group to discuss, but your point is well taken. There were no hemostatic data. All of this was done at a time when bleeding times and CCIs were fine, "Happy Valley," but we are not there anymore. So, your point is correct.
DR. GOLDSMITH: You talked about decreased wastage post-pooling, that these platelet packs could be returned from the OR. Could you comment on the storage of the platelet packs in the OR, what impact that might have on the viability of the platelets?
DR. SNYDER: I could. It is a problem when platelets go out. We would not allow them to come back the next morning. We would have to think of a new paradigm. If we gave out a platelet pool we would have to track that. Most likely it would go out with some form of temperature monitor or in some type of a container so we could have a little comfort level. We would have to in-service the physicians and nurses that this was not something--although some people probably hand out units of random donor platelets that are not pooled and they let them pool them at the point of care, we don't do that. Those come back as well. The same thing would apply to red cells. You would just have to make sure that the quality systems were in place in the operating room. But your point is extremely well taken. You would also have to be sure that it wasn't entered in some fashion and came back. But those are all concerns that I would love to address if we had a chance to have that type of a product.
DR. SCHMIDT: I have some questions of Dr. Vostal. Are we going back to that?
DR. NELSON: Is it a question? I mean, we are going to continue discussion. Final comment?
DR. ALLEN: Just a question with regard to the filtration. At the present time we have an infusion period within four hours. Is it normally done at the time of pooling? You pool and then run it through the filter and then release it within a few hours?
DR. SNYDER: Before I answer that let me say that it is not filtration per se; it is leukoreduction. So, process leukoreduction should be, in my perspective, as good as filtration leukoreduction. That having been said, it depends. Yale is 100 percent leukoreduced so if we pool it has already been leukoreduced. Others may want to leukoreduce in the lab, may leukoreduce as a pool because it is less expensive to use one filter than to have to use three or four if you are pooling three or four units. So, it depends on the facility and what their practices are.
DR. ALLEN: If the storage period for pooled platelet packs were to be extended to, let's say, your proposed 72 hours, you would expect that the leukoreduction filtration might be done at the time of initial pooling or just pre release?
DR. SNYDER: I personally would recommend leukoreduction at the time of preparation up front. I wasn't recommending 72; I said at least 72. I think five or seven would be okay as well. But I would not recommend post-storage leukoreduction. I think your best benefit, from the comments that were made by the gentleman before, is that pre-storage is best because it prevents cytokine generation; it prevents microparticles, etc.
DR. ALLEN: Thank you.
DR. NELSON: I think we have already discussed that in a previous BPAC meeting.
DR. SIEGEL: Yes, we actually have published a draft guidance on leukocyte reduction. We encourage routine leukocyte reduction but cannot mandate it without rule-making. Be that as it may, for labeling products that are leukocyte reduced we strongly encourage pre-storage leukocyte reduction for the reasons that are being reviewed.
DR. NELSON: Let's take a half hour break, until eleven o'clock.
DR. NELSON: Back to platelets. Next discussion is by Dr. Pietersz to present the European experience with extended storage of platelet pools.
European Experience with Extended
Storage of Platelet Pools
DR. PIETERSZ: Mr. Chairman, Madam Secretary, members of the Committee, I want to thank you for your invitation to be here and to share with you some data on buffy coat platelets. It is a special pleasure because the godfather of the buffy coat was Peter Prinz and later Hans Lowells from the Netherlands, and also Klaus Hoekmann from Sweden has been working a lot with the buffy coat development.
I want to share with you the names of my collaborators in this huge investigation we did, Peter van der Meer is a post-doc. senior scientist. He did a lot of work on the in vitro studies of the buffy coat platelets. Margrit Dijkstra is a Ph.D. student and Hank Reesink, as many of you know, is our driver or research in the Blood Bank of Amsterdam.
The Netherlands is a small spot on the map of Europe, and since '98 we have had a national foundation of the blood supply, named Sanquin. We have 60 million inhabitants and from February, 2002 we have had four blood bank regions and we have less than one million donations, 50,000 whole blood and 125,000 plasma apheresis. From these, 250,000 are used for random platelets and only five to ten percent of the 250,000 are apheresis platelets.
At this moment we have nine product sites but we will reduce that to four before 2004. We have four test labs including NAT testing and at this moment we still have 13 distribution sites because the traffic is very high.
Sanquin has national requirements and I have listed a few of them here. We have bacterial screening of all platelet concentrates or products from November, 2001. We have universal leukoreduction of cellular products from January, 2002. Sanquin also decided to have leukoreduced plasma and we have been collecting it since July 2002, and it should be issued as donor re-tested from April, 2003. There is a blood component guideline and I am the chairman of this committee. In January, 2003 we had the sixth edition.
We have component preparation. In the whole of the Netherlands we collect 500 mL of whole blood, plus/mins 50 mL. Our maximum collection time for platelet preparation is 12 minutes. We cool the whole blood within two hours with cooling elements to approximately 22 degrees C. and then we store it at ambient temperature. Ambient temperature is defined as 18-25 degrees. You can make the separation of whole blood within four hours of collection, but usually we store it overnight for 12-22 hours.
Next, whole blood is given a hard spin centrifugation, about 30,000 g minutes. Then we have separation into plasma, buffy coat and red cells, and the red cells are suspended in SAGM solution, comparable with ADSOL, and we filter the red cells with an integrated filter.
For the separation we use automated equipment and we have programmed this equipment to have specific buffy coat specifications. The volume should be 50 mL, plus/minus 5 mL, a hematocrit of approximately 0.40. If we make platelets from them, we should have at least 75 X 109 platelets in the buffy coat.
For the buffy coat pooled preparation, we use five ABO identical buffy coats and one plasma, or you could use platelet additive solution. In many Scandinavian countries they use additive solution, and we have used plasma and later solution but now we are back on plasma.
We use an integrated set bag system and then we register all the units, the original donation units, in a computer and link it to a pool number for traceability. It is very important that the whole set should be connected. Then we register in a computer.
Then we transfer the content of the buffy coats into a pool bag. We add the plasma and give the whole pool bag a soft spin centrifugation and, again using automated equipment, we express the platelet-rich plasma through a leukoreduction filter into storage bags.
I will now show you this handling in pictures. Here you see an integrated set. This set is from Terumo. The set has six leads, pig tails, to which buffy coats can be connected with sterile connection and also the plasma. Then you have the pull-back here. The pull-back is connected to a leukoreduction filter and this is, again, connected to a platelet storage bag and we have a small sample bag and this has a special adaptor for later bacterial culturing.
We make the sterile connections so all the leads are checked. When the sterile connection is made we check all the connections.
Then the content of the buffy coat is just drained by gravity into the pool bag and also the plasma.
After soft spin centrifugation we have the centrifuge back on the automate. It is Compomat G4 from Cerus. We express the platelet plasma through the leukoreduction filter into the platelet storage bag, which is up here. Finally, we let the content of the filter drain into the platelet storage bag because most platelets are in the last milliliters of plasma so we want to harvest these as well.
You can see here that this automate helps us in timing of closing the tubing because if it sees the red cells, as you can see here, the sealing is closed and then the expression is terminated.
Here you see a platelet pool. It is labeled with ISBT code 128. We have had this bar coding already since '98.
The platelet pools are stored on a horizontal agitator in a climate cabinet of 22-24 degrees for seven days.
Some advice for buffy coat pool are that the volume of the pool bag is of importance. The content is about 550 mL. So, you have a pretty filled bag. The hematocrit of the pool is about 0.2. Depending on the viscosity, you either have buffy coats in plasma or buffy coats in solution. So, that is a big difference. You centrifuge at about 1,000 g. For us, we have a requirement that the yield of platelets following filtration, as you always lose some platelets at filtration, should be higher than 80 percent.
These are quality control data of 2001. We have made over 10,000 pools, about 70 200 pools per working day, four working days per week. You see that the volume is approximately 312 mL and the requirements are 150-400 mL. For platelets we have a minimum requirement of 250, and you can see that the number of platelets, even without the 25 mL sample that we have taken off, is 364. Leukocyte counts should be below one million and you can see that it is far lower than that.
The pH is measured at the day after expiry, on day eight. This is the quality control and you see here that the pH measured at 37 degrees is still very acceptable.
Here you see some differences between buffy coat pools made in plasma compared to buffy coat pools made in PAS-II. For about a year and a half we have used plasma and then we went back to PAS-II and we are now again on plasma. We like to switch sometimes. But you can see that the yield of platelets in the plasma pools is higher than in the PAS-II. You can also see that the number of pools having platelet number lower than 240 is very low. Over 99 percent contains more than 250 X 109 platelets. For leukoreduction it doesn't matter whether you have plasma or PAS-II as the storage medium.
Anderson showed that the yield of platelets in buffy coats is slightly higher than from PLP and almost equivalent to an apheresis unit, just to show you differences.
Now I would like to discuss storage with you. Dr. Snyder already showed that the pH is very important for platelet storage. If you compare the pH and the platelet morphology, at pH 7.4 platelets will have disc-like morphology. They are happy platelets. At about 6.8 you will have spheres but this is still reversible to discs. In the Netherlands we have decided that for our requirements platelet concentrates should have at the end of storage a pH between 6.8 and 7.4 measured at 37 degrees. There is a difference between 22 and 37 degrees but we have this requirement. Of course, you can go lower with the pH but then you will have balloon formation. The morphology will sort of resemble a balloon. This is not reversible anymore so these platelets are already not happy. If the pH is even lower they will die.
The drop in pH is caused by CO2 formation if enough oxygen is present, and without oxygen you will have lactic acid formation and the pH will drop even more quickly. For storage of pools it is very important what kind of storage medium you have because if the storage medium has a buffer capacity, then you will have a slower drop in pH. For instance, plasma has a high buffer capacity. Additive solution does not have a buffer capacity. The platelet concentration is very important, the number of leukocytes and, of course, the gas permeability of the container and the breathing surface is very important. I will now show you some examples.
Peter van der Meer did an experiment where he made a large pool of buffy coat platelets in additive solution. These are containers of various sizes and also of various materials. Here is 1.3 mL PL2410 bag, Polyolefin bags 1.0 L and 1.5 L, 1.5 L CLX bag and a 1.0 L bag of DnDP-PVC. You see that during the storage for nine days almost all bags were appropriate in considering the pH. Only one bag had a quick drop in pH after day five. This is the swirling effect. We have a swirling score from 1-3 so 3 is the best swirling score and 1 is the score where we consider the platelets not healthy anymore.
You see the same picture and the pH. This one bag showed a big drop in pH and you see here that at 6.8 the line is crossed at about five-something days.
From these experiments we concluded that it is very important to validate the size and the type of container you store the platelets in if they are stored as a pool. But there are many bags that are very acceptable to store the pooled platelets in.
Another thing that is of great importance is the concentration of the platelets. Peter van de Meer took all the data we had. The yellow ones are the platelet concentrates in plasma stored as a pool. The blue ones are platelet concentrates stored in PAS-II. You can see that where this line is crossing the 6.8 pH line the concentration is about 1.4, or somewhat higher.
So, the conclusion is that you can easily store platelet concentrates at a concentration between 0.8 and 1.4 in all the bags I just showed, whereas for additive solution it is very important that the concentration of platelets is much lower, only 1.1. If you store as a pool it is very important that your storage medium equals these concentrations, otherwise, if you have higher concentrations you will more easily have a drop in pH.
I would like to switch now to some results of our bacterial screening. It was conducted as of November 1 but we started already in June, 2001. I show some of the results of one year follow-up of the components involved in the screening of leukoreduced platelet concentrates stored in plasma for up to seven days.
You saw the adaptor on the sample bag. We inoculate aerobic and anaerobic culture bottle from BacT/Alert from Biomerieux equipment. It is very easily done. With the special adaptor you can very elegantly fill the bottles in an air flow cabinet, and we collect in each bottle 10 mL. Here it is marked how many milliliters are collected in a bottle so it is very easy, and we have a sample of 10 mL. So, our total sample from the pool of platelets is 25 mL that we use for culturing and QC data.
We incubate the bottles in BacT/Alert for seven days or until a positive signal comes up. The positive signal is given by CO2 development. The whole system is linked to a computer so if a positive signal comes up, then the monitor will flash and you will know exactly which bottle is positive. If that happens, we have confirmation of the bottles in a microbiology lab and I think it is very important for discussion that our definition of confirmation is that bacteria is growing in the culture bottle.
In the meantime, the issue of leukoreduced platelet concentrates is negative to date so we always check in the computer whether the concentrate has a negative signal. If the signal is positive, of course we block the pool and also the respective red cells in the center. Because our hospitals have their own inventories of platelet concentrates, we also notify the hospitals and they also block the pools in inventory and eventually the red cell concentrates. All the blocked components are recalled, and if there has already been a transfusion we have a request for information about the transfusion.
We cultured about 9000 platelet concentrates. In 81 cases we had a positive signal and there was no confirmation in the culture bottles in five cases, whereas we found bacteria in 76. So, we had a positive rate in the pools of 0.9 percent and when we looked at the positive ones during the months we saw a slight variation and especially, for the Netherlands, in warm months of July and August we saw a higher incidence of positive signals.
These are the detected bacteria. Most of them are skin bacteria, except for this one. What was of interest is that, for example, the Bacillus species grew within 24 hours. So, about 50 percent of the bacteria grew within 24 hours of inoculation, whereas the Propioni was growing late, after four to five days of inoculation.
From about 9000 platelet pools, 81 gave a positive signal and 76 were confirmed positive in the culture bottles. Predominantly skin flora was cultured and about 50 percent of the cultures became positive within 24 hours of inoculation, whereas the other 50 percent after four to five days of incubation. Most of these contained Propioni and we had no septic episodes mentioned by the hospitals.
I would like to share with you now some data of our clinical investigation of the five to seven day stored platelet pools. We included outpatients with hemato-oncological diseases. Exclusion criteria were serious bleeding, WHO grade 3 or 4. So, bleeding that required transfusion of red cells. The pre-transfusion counts should be lower than 20 X 109/L. So an exclusion criterion were platelet counts higher than that, and refractoriness defined as two consecutive unsuccessful transfusions.
We calculated the count increments and corrected count increments.
We found that the data showed that 341 out of 349 transfusions of platelets of all storage data were successful. If a successful transfusion was defined as CI higher than 10 or a CCI higher than 7.5 we found 95 percent successful transfusions. If we compared the five-day old and the seven-day old platelets, you can see here that there was no difference in unsuccessful transfusions.
If we compared the transfusions given to the same patient and then five-day old or seven-day old, you can see here that there is a difference in count increment. The seven-day old was slightly lower than the five-day old platelets but, as yet, we don't know what kind of difference is acceptable. If you say that a 20 percent difference is acceptable because, of course, older platelets will have lower increments, then this is still an acceptable difference.
When you calculated the corrected count increment and also took the transfusions given to the same patient of five- and seven-day old stored platelets, we had no difference but the power was only 31 percent. So, we concluded that we need more data.
We concluded from this clinical study that is still going on that 95 percent of the transfusions were successful for corrected increment and the CCI; that transfusion of leukoreduced platelets five- and seven-day old in the same patient had slightly lower CIs but that much more data are required in each arm for confident proof.
My final conclusions are that leukoreduced platelet pools can reliably be produced from buffy coats and integrated systems. I think that for sterility it is important that you have an integrated system. If you decide to pool PRP platelets, I think you can easily use the same system.
In vitro data are sufficient. Dr. Snyder also showed a lot of data and from Europe we also have other data but I could not show you all of it to allow seven-day storage in appropriate storage bags. So, validation of the storage bags is very important. Clinical evaluation of these buffy coat platelets show 95 percent sufficient increments of seven-day old platelets. Bacterial screening of all PC is feasible. It does not prolong the issuing of the platelets, and we found 0.9 percent of our pools positive but, again, this is the culture bottles that gave the confirmation and I do not know what the next speaker will say about that. Fifty percent was positive within 24 hours so most of these pools could be tackled within the blood center. Thank you for your attention.
DR. NELSON: Thanks, Dr. Pietersz for nice data. Questions? Were you able to interdict the platelet pools that had growth within 24 hours? It seemed like that would be feasible.
DR. PIETERSZ: Yes, they were still in the blood center.
DR. NELSON: So there were no adverse events from those positive cultures?
DR. PIETERSZ: No, and we think, but maybe the next speaker will explain that in more detail, that those were the most dangerous bacteria.
DR. NELSON: So, in those where there was delayed growth you showed no febrile reactions or adverse events in the recipients? Is that correct?
DR. PIETERSZ: We did not have any records of septic episodes.
DR. LEW: I think when I was looking at the literature they suggested that typically only one bag in, I think, a thousand or so is usually infected, and your data showed almost one percent or one in one hundred. Do you believe it is because you pooled first and then tested? Also, it sounded like the bug that most often grew is a slow grower. Is that responsible for the 50 percent rate on day four and five versus the idea that you did a dilution so it took more days to grow?
DR. PIETERSZ: I think what is important is the definition of the confirmation. We had a meeting in Dublin last week, and if you say that you take a sample from the pool, you inoculate that in culture bottles, and if you would say that confirmation is repeat culture from the same sample, then we only had like 0.009 positives. But we defined confirmation as growth of bacteria in a culture bottle. So, that might be a difference in definitions because otherwise we would also have an incidence of 1/1,000, or even less.
DR. NELSON: Thank you. Next is Dr. Mark Brecher, from University of North Carolina, bacterial detection in platelet products.
Bacterial Detection in Platelet Products
DR. BRECHER: Thank you. It is a pleasure to be here and I welcome the chance to talk about one of my favorite subjects, bacteria.
Similar to Dr. Snyder, in the interest of full disclosure I just want to say that I receive research support and am a consultant for certain advisory panels for a variety of companies, including Biomerieux, Pall, Baxter, and several others. If they don't support my research, please see me after this meeting.
What I wanted to do was to give a general overview of the whole question of bacterial detection, and I think we need to put this into some perspective. The rate of bacterial contamination is about 1/1,000 to 1/2,000 bags that are bacterially contaminated. When we talk about the various viruses, we have done a wonderful job over the years of bringing down the risk of HIV from about one percent to about one in two million. HCV in the early to mid-'80s was about half a percent, and it is down to about one in one and a half million and hepatitis B. But, clearly, the risk of bacterial contamination of platelets is orders of magnitude greater than the risk of viral transmission.
Looking over the literature, I tried to figure out exactly how big a problem we are really dealing with and I keyed in on the United States. In the United States there are roughly four million platelet bags transfused per year. That is, one million single donor apheresis packs and three million random donor platelets made from whole blood concentrates. So, approximately three-quarters of the doses that we are handing out are apheresis platelets and 25 percent are pools of random platelets.
Now, if there are 1/1,000 to 1/2,000 bacterially contaminated bags, that means that every year we are handing out 2,000 to 4,000 contaminated bags of platelets. The literature suggests that if you transfuse one of these bags, roughly one-quarter to one-sixth will have clinical symptoms. We are making about 333 to 1,000 patients sick per year. The literature worldwide suggests that of those who are symptomatic. So, roughly one-fifth to one-third will die so that means 67 to 333 deaths per year or a fatality rate per unit of approximately 1/12,000 to 1/60,000.
Are these numbers real? I think the best data comes from Paul Ness and his group at Johns Hopkins were they found that the risk of dying from a pool of random platelets was 1/17,000 and the risk for an apheresis fatality was 1/61,000. Of course, they only know about the cases that were reported back to them so the real numbers may even be higher. So, I think that these numbers approximate the real risk in this country from bacterial contamination of platelets.
The bacteria that cause fatalities represent a broad spectrum of bacteria. Staph. aureus and Klebsiella are probably the two biggest ones but Serratia is not very far behind. Roughly 59 percent are gram negatives, 59.7. So, roughly 60 percent of the deaths are due to gram negative organisms, of which the majority are Enterobacteriaceae, Escherichia, Salmonella, Proteus, Klebsiella and E. coli. This is quite distinct from what you actually find when you culture units. When you culture units roughly two-thirds are gram positive organisms but when you look at what actually kills you the majority are gram negative organisms.
So, we have to come up with a system that will interdict all of these bacteria if we are really going to address the problem. Where do these bacteria come from? Most of the gram positives are thought to come from the skin. When we sterilize the skin we really don't sterilize it. All we really do is bacterial load reduction and there are still some bacteria left on the skin that will come up the needle. In addition, the needle goes through the subcutaneous tissues, cores our some of the subcutaneous tissues, sebaceous glands, hair follicles, etc. where the iodine solutions don't reach. So, some of the bacteria are coming up that way. Most of the gram negatives are thought to come from a donor who has a transient asymptomatic bacteremia. Those are the two major sources of bacterial contamination of platelets.
When we are talking about bacterial detection and the detection of reactions to platelets, it is all about timing. This is a series of cases that were found at the NIH Clinical Center, back in the early '70s. During their epidemiologic surveillance they found that there was a series of Salmonella cholerae-suis cases in the hospital. This led to a major epidemiologic search trying to identify the source of the Salmonella. To make a long story short, they found that they all traced back to the fact that they had all had platelet transfusions and all of them had received platelet transfusions from the same donor.
Further investigation showed that that donor had osteomyelitis so every time he came in to donate platelets he was shedding Salmonella. The interesting thing about this case is how long did it take for the patient to become sick after they received the platelets? There was quite a range. It went from 5 days to 12 days. The average incubation time before the patient became symptomatic was 8.6 days.
So, it is no surprise that these cases were not linked to platelet transfusion. This is sort of a recurring theme we see in the literature. Many of the patients become sick sometime after the transfusion and they are not linked back to the original platelet transfusion unless someone comes looking, trying to figure out whatever happened to other components made from another bacterially contaminated product. In this case there were seven cases and one patient died and two patients had recurrent Salmonella episodes.
When you do inoculation studies you put bacteria in a bag and you look and see how fast they grow, This is looking from a study that we did several years ago of 120 inoculated platelet units with Staph. epi., Staph. aureus and Pseudomonas. We have done similar studies with a variety of other bacteria, the vast majority will become culture positive within one to two days. However, with Staph. epi. you often see that sometimes you don't detect bacteria in the bag until quite late. Staph. epi. can be bit of a slow grower.
This is data from the Holland Labs, the American Red Cross, Steve Wagner, where they put E. coli and Staph. epi. into bags. They inoculated with a low concentration of 0.1 organisms per mL and then they put a relatively small volume into a bacterial culture bottle, 0.5 mL. Normally when you put a volume in a culture bottle it is 4-10 mL. With E. coli by 24 hours they picked up 100 percent but it required greater than 24 hours to pick up all the Staph. epi. because sometimes it takes time for these bugs to grow in the bag. Unlike viruses, where you can detect viruses in a sample right at the time of donation, you have to allow these bacteria to grow before a small sample taken from the bag will allow you to detect the bacteria.
This is similar data from the Canadian Red Cross, back when there was a Canadian Red Cross. On day one, which is the day on which the platelets were collected--I would have called it day zero--they found a contamination rate in 16,000 platelets of 0.02 percent. However, they came back two days later and they found a contamination rate of 0.07 percent. At that point there were only 10,000 platelets left to retest. Again, it illustrates that you have to let some time go by before you do your sampling if you are going to pick up all the bacteria.
One point I wanted to make is that in the United States we have two platelet products. We have single donor apheresis platelets, shown here, and then we have pools, which have already been discussed quite a bit so far this morning. What we really have is a two-tiered safety system. Since the chance of getting a bacterially contaminated platelet is a function of the number of donors and the number of needle sticks, apheresis platelets are intrinsically safe. Data from Johns Hopkins, where they have made a conscious effort to switch to 100 percent apheresis products, they found that the risk of getting a fatal transfusion or just becoming septic from platelet transfusion was five- to six-fold higher for pooled random platelets. So, we really have a two-tiered safety system in the United States and I hope we don't make this worse.
The shelf life of platelets is limited to five days because there is a fear that bacteria are going to grow over time. At one time platelets were licensed out to seven days, in the early '80s. When you do inoculation studies at a variety of relatively low concentrations, what you find is that usually by day three, sometimes day four, the bacteria are already on the plateau of growth. Usually when bacteria grow it doesn't take five days to reach plateau. Usually they are there by day three. With some organisms, like Bacillus, usually within 24 hours they are on plateau growth. So, a day one or a day two platelet can be quite dangerous.
In the data from the SHOT study from England, the Serious Hazard of Transfusion study, they had five deaths in the first couple of years of their surveillance study. Three of the five deaths occurred from platelets that were days one to three. So, early platelets can be quite dangerous as well.
This is data from our lab looking at Staph. epi. and shows that growth with Staph. epi. can be much more variable and can be quite slow. It seems to be somewhat of a function of the initial concentration that is put into the bag, unlike most of the other bacteria where it doesn't really seem to make much difference what the initial concentration is, but Staph. epi. is a problem.
This is jut to remind me to emphasize that you cannot test the whole bag for bacteria. So, you have to let some time go by, and we are running out of time here--April, 2003.
There have been a variety of detection schemes that have been explored. Fortunately, many of these have played from my laboratory over the years. Many are low tech, such as looking at swirling, versus high tech, using antibiotics as probes. I am not going to talk about those because they never really made it to market. They are not available yet. So, I am just going to talk about what it is that we can do today.
We can do bacterial staining. This can be a Gram stain. It can be a Wright stain like they use in hematology labs. Most hospitals have these automated Wright stainers. You put a slide on a little conveyor belt and out it comes on the other side, a beautiful slide. And, we don't really care whether it is gram positive or gram negative; we just want to see it. However, with the Wright or Gram stain--this is data from Lee Bland who used to be with the CDC--you find that the pickup rate is only about 106 CFUs per mL, so a million bacteria per mL. You could go to acridine orange and fluorescent microscoping and that gets to be a log better, but it is not that sensitive a system and looking at stains is very subjective and you can have a high false-positive rate.
An alternative relatively simple system that was initially described from my lab was to use multi-reagent strips, urine dip-sticks. We keyed in on two markers that could detect bacteria. One, we reasoned that as the bacteria grew they would consume glucose. So, we were actually looking for negative glucose. Basically, as they grow bacteria consume all the glucose. And, we looked for a drop in pH. We were looking basically for an orange pH, not green, a pH of 6.5 or lower. Actually, there is some bias in these strips when you use them on platelets as opposed to urine.
This is a paper where we published some of these pictures. This is looking at platelets that were inoculated on day zero. These strips were made on day three. You can see that with the Klebs. pneumoniae the glucose is blue; the pH is orange. Staph. aureus, blue and orange. On day three both of these bags were at 107 CFUs per mL, basically on plateau. However, it doesn't pick them all up.
Here is S. marcescens. The glucose is green, as is the pH. This was at 103 CFUs per mL. So, data from our laboratory and from other laboratories like Steve Wagner's lab, suggest that the dip-sticks are sensitive for about 107 CFUs per mL and in some cases will pick things up down to 104, 105. But it is a very easy test to do. The strips are very inexpensive. But they have not really received wide use.
There was one paper from M.D. Anderson Cancer Center that was published in Transfusion last year. They looked at roughly 3,100 random platelets. They used the dip-sticks. They found two that were contaminated and were able to interdict those units so they were not transfused and very likely saved two patients' lives. However, there was a relatively high false-positive rate in that 28 units failed the dip-sticks. Either the pH was too low or the glucose was too low. Those were pooled. You might ask were those units any good anyway if the pH was low or the glucose was low. Maybe it is a good quality control measure, but a simple technique like this can interdict contaminated units.
We already heard a bit about the BacT/Alert system from Biomerieux. This is one of two systems that are currently approved by the FDA for in-run quality control of platelets. The BacT/Alert 3D, as you have already heard, has a color change. It goes from green, which is good, to yellow, which is caution. BacT/Alert 3D is a newer machine than the one that you saw from the Netherlands. Shauna Hayes is my research tech. Basically, what happens is that these bottles are checked roughly every ten minutes so it is almost real time. They look are reflectance of light off the color sensor. Not only does it look at what color is it absolutely, but it is hooked to a computer and keeps track of what is the rate of change in the color. So, the computer will pick up bottles that are positive before the human eye will notice the difference in color. So, it is a relatively quick system.
We have validated the system in our lab, looking at 15 different organisms that are known to contaminate platelets. What we found, using a variety of the bottles that were available, is that when we targeted inoculations at 10 CFUs per mL--the actual mean was 10.9 but, in fact, many of them were less than 1 CFU per mL--is that for the vast majority of these organisms, we could pick them up generally roughly in 12 hours and, on the outside, about 25 or 26 hours, with the exception of P. acnes, Proprionabacterium acnes, which is an anaerobic organism that is a slow grower and is generally thought to have questionable clinical significance, particularly in the context of platelets. But for the aerobes the system does extremely well.
Getting into the bags, sterility has been a problem. Sterile connection devices are the way you access these bags and there have been a variety of sampling devices that people have been exploring in this country. This is one we have been using. This is similar to the system you have already seen from the Netherlands. This is a system we are having custom-made for our lab where you sterile connect this on and there is a syringe. You just pull up the volume you want. There is a needle on this end and you put that into the bottles and you put in the exact volume that you want.
What will it take to use a system like this to extend the shelf life of platelets? Assuming that the in vitro function and in vivo survival is fine, there is a question of whether an early culture is predictive of a late culture. That has already been alluded to this morning. At the last BPAC meeting, it was discussed what it would take to do a study.
We have been doing a pilot study for the last year. In our first 12 months we sampled roughly 2,400 apheresis platelets. What we found was that we had one collection, which was a triple collection, that had Staph. epi. The machine triggered at 14 hours. We were able to interdict all three of those units so they did not go to a patient.
We also had two collections that were doubles, so four bags, that came up with Proprionabacterium acnes on day six of culture. We were setting up our cultures on day two of storage. So, those units had already been transfused several days before. When we went back to talk to the clinicians, there were no ill effects attributable to the P. acnes in those patients, which gets to one of the questions that was asked before.
In terms of a false-positive rate, we were sampling every bag with an aerobic bottle and an anaerobic bottle on two occasions. We basically inoculated 9600 bottles during this time period. We had one bottle that was contaminated by P. acnes presumably from my skin. I inoculated that bottle. I take full responsibility for that one. So, our false-positive rate was on the order of about 1/10,000 from inoculation. The way we proved that, we kept a sample bag on every bag and we only said it was a real contamination if we could either reculture the original bag or the sample and find the same organism in that bag.
In theory, in terms of contamination I think you can get a rate as low as 1/10,000. That is a lot lower than the data we heard about from the Netherlands. Jim AuBuchon has been running a similar study at Dartmouth and his contamination rate has been running 0.5 to 0.6 percent, which is much more similar to what they see in the Netherlands. That is perhaps what you see out in the real world where you have a lot of technologists doing it.
So, what do we need to get this licensed? Well, the FDA says they don't want to see inoculation studies. They want to see real-world data. However, in my mind, where do the bacteria come from? They come from the skin. They come from trans-bacteremics.
I don't know that inoculating bacteria into platelet bags is that much different so I am not sure that we really need real-world data. But if we were to do real-world data, the biostatisticians would tell you that they want to see 50 to 100 real events. Since we have a contamination rate of maybe 1/1,000 we would have to study 50,000 to 100,000 platelet bags to detect 50 to 100 real events. That sounds quite daunting but I think it can be done and we are having some discussions with some of the large blood collectors to see if we might be able to pull off a study like that in the next year or two, although I really question whether we really need to do that. That is my opinion.
The other system that is currently approved for detection of bacteria for in-run quality control is the Pall bacterial detection system BDS. This is a small pouch that is sterile connected to the platelet bag. You push over 6 mL as opposed to the BacT/Alert where we were using 8 mL. It goes through a filter which filters out the white cells and the platelets but lets the bacteria pass. It is different depending on what organism you are looking at but, on average, 50 percent of the bacteria will pass the filter. It then goes into a little pouch that holds 2 mL, which has an SPS tablet which will enhance growth in this bag. Then you put that at 35 degrees centigrade in an incubator and let it sit for 24 hours. As the bacteria grow, they will consume oxygen. As opposed to the color changes that you saw with the BacT/Alert which was looking at CO2 and pH changes, this looks at consumption of oxygen. So, you measure the head space gas, and you look to see whether it is a pass or fail.
Although there haven't been papers published on this, abstract form presentations at meetings suggest that this is sensitive down in the range of 10-100 CFUs per mL. This system is only good for aerobic organisms, not for anaerobic organisms. It has the advantage that it is a much more closed system. You don't have a needle waving in the air in the laminar flow hood so there is potential for less false-positive reactions.
What does the future hold? Teh future usually comes from the past. So, let's go to the next slide.
Where we are is here. This is a quote, one of the recommendations from the Institute of Medicine report on HIV in the blood supply where they said that the perfect should not be the enemy of the good. The implementation of partial solutions that have little risk of causing harm should be encouraged. That is where I think we currently are with bacterial contamination of blood products. We are at the point where we are starting to phase in partial solutions. They are not perfect but they are pretty good.
Here is a sage fellow, Ed Snyder. There have been three major FDA workshops that have dealt with bacterial contamination of blood products. Many of us in the field have felt that we need to do something about this. I thought Ed, in his summary comments from this meeting, said it better than anybody else: The imperative is to act so you don't have to explain yourself on Night Line. Regulation is necessary to achieve the goal. Nothing says I care like a page of 483s. Those are the deficiency forms you get from an FDA inspection. When all else fails, do something. Give us a mandate and we will do the rest.
I think it is key, we have to have a mandate from somebody, be it the FDA, AABB or somebody, or we won't be able to convince the hospital administrators to let us do something.
Following the third meeting last summer, which was principally on pathogen reduction but it certainly found itself in bacterial contamination waters quite a bit, a group of the moderators and speakers got together and wrote an open letter to the blood banking community. What we said in this letter was that pathogen reduction has a lot of promise but it is not going to be here soon. Meanwhile bacterial contamination is a major problem in this country and we need to do something now. So, we urge the blood collection community to begin implementing bacterial detection strategies now. This was a letter from myself, Jim AuBuchon, Paul Ness, Roselyn Toby and Bill Blackman.
This letter actually generated some very interesting responses. Actually, it was posted on a couple of web sites. Some people accused us of moral blackmail and asked us about what conflicts of interest we had with the various companies. Like Ed, I have no equity interest in any company.
So, things have been changing in recent months. The AABB has been considering a standard for the past six months. The original proposed standard was that the blood bank or transfusion service shall have a method to test for bacterial contamination of all platelet components. This wording has been changed a bit but the essence is still here. The standard has been approved as of last week, and will be effective in this country on March 1, 2004. So, essentially we have a year to gear up to do bacterial detection of platelets.
Similarly, the College of American Pathologists, CAP, and their laboratory accreditation program also now has phase I deficiency, which is in many cases a recommendation, does the laboratory have a system to detect the presence of bacteria in platelet components? They want you to look at all of the platelets. So, things are changing. There is a mandate and we are moving towards bacterial detection in this country.
What a lot of us want to see happen is that we want to see the shelf life go back out to seven days. This is just a timeline. In '82 platelets were extended to five days from three; in '83, from five to seven. But because of several reports of bacterial contamination in older platelets, it was brought back down to five days. We think we can probably get it back out to seven days if we have a bacterial detection step somewhere in the system.
We have already heard that bacterial detection is being used in Europe. There are several countries now that routinely use a culture system for all their platelets. In addition, many of these countries are now extending platelets out to seven days, either countries or institutions--Denmark, Netherlands, Yugoslavia and United Kingdom. There have been reports of extending platelets to seven days if you do a bacterial detection step, usually on day one or two of storage.
The other interesting thing is that it is very cost effective. We throw away about ten percent of the platelets in this country because of the outdate. If we could get an extra two days on the shelf life of platelets we would drop the outdate rate substantially and it would easily pay for itself. Jim AuBuchon, who I know has spoken to this committee before, is generally the guru of cost effective analysis as it applies to transfusion medicine. In virtually every cost effective analysis he has done--leukoreduction, p24, NAT testing--he always concludes it is not cost effective to do something and yet we wind up doing it anyway. In this case, he finally concludes that this is very cost effective. It is not even cost neutral; it would pay for itself. If we didn't outdate all those units, it would pay for all the culturing. Yet, we are very slowly walking toward that.
So, what are we going to do with random platelets? Well, we have a bit of a conundrum here. We have these sensitive culture systems that are usable for apheresis platelets but don't lend themselves to random platelets because of the volume that must be withdrawn and the cost of the systems. It is about $25 a culture with either of the two systems that are licensed.
What we are currently faced with is that if we are going to do a bacterial detection step for random platelets, unless we can pre-pool, we are going to use systems that are less optimal. We are going to use the dip-sticks. We are going to use Gram stains or Wright stains which are not the best systems.
In addition, we have heard some data that said, yes, there is a risk that you have greater bacterial growth in a pooled random platelet if one of those platelets is contaminated but it is not a log difference in total dose. I don't think 106 bugs per mL versus 107 is going to make a difference in the patient outcome.
In addition, what also hasn't been said is that in several of these pooling experiments, when you pool several platelets together often the bacteria didn't grow because one of the donors in that pool had antibodies that basically killed the bacteria. So, in some cases pooling may be better in preventing bacterial contamination of platelets.
Nevertheless, unless we have pooling people are trying to figure out other ways of doing cultures so that we can try to culture the platelets. One system is this system which looks very complicated, whereby you would take approximately 2 mL from each random bag, bring that up into a syringe and then you would inoculate it into the bottles. You would leave all the original bags connected to the setup, ship it all out to the hospital and then, at the time that you are ready to pool, you pool into this common bag, here. So, it is pre-pooling that is not pre-pooling. You pre-pool just the sample but it is a lot of manipulation.
Alternatively, Pall has a system where you put a little autostat filter--both of these systems require leukocyte, or at least they were approved with leukocyte reduced platelets where you take your platelet-rich plasma, you can pass it through here, make it into platelets and then sample the system. You know, this is a lot of work for each little random platelet and it is going to be expensive. So, we are struggling with what we can do with platelets.
When the AABB standard goes into effect we are going to take a two-tiered system of safety in regard to bacterial contamination and we are going to still have a two-tiered system because we are going to be doing sensitive testing on apheresis platelets and probably much less sensitive testing for random platelets. That makes a lot of us very uncomfortable.
DR. NELSON: Thanks, Dr. Brecher. Yes?
DR. LEW: Back to how often you actually get a positive, your definition was different from Dr. Pietersz. Dr. Pietersz said up front, you know, in the real world if you get a positive you can't wait around a couple of days for a second culture to come back. So, they usually hold it and then, you know, not transfuse. So, if you were using her same definition what would be your rate? Because that is the real world, if it goes out you can't wait for the false positive to come in necessarily.
DR. BRECHER: Right. One thing you have to realize is when people toss out these numbers of 1/1,000 and 1/2,000 bacteria contaminated, those studies were all done with aerobic culture systems only. Now, all of a sudden, we are looking at anaerobic systems as well.
In our experience with 2,400 platelets, we would have had seven that were truly contaminated that we detected and we would have had two that triggered on the machine that were not really contaminated, one which we contaminated ourselves with P. acnes and the other was a machine false positive. So, we are talking about roughly 10/2,400 so it is well under one percent. It is a low rate.
For the real world I think Jim AuBuchon's data is probably better. He was seeing about 0.5 to 0.6 percent that you may have to throw away. But even when you figure that into the scheme of things, it would still be considered cost effective.
DR. LEW: If I could just follow-up, do you think though that by pooling early--I don't care what the number is, even it is just 1/1,000--but if you pooled early then you would expect that to actually go a little bit higher.
DR. BRECHER: Right, yes. That was clearly part of what was driving the numbers that Dr. Pietersz presented because she was pooling. But I really don't know what was true positive and what was false positive from her data because they didn't have a follow-up sample.
DR. NELSON: Her data was expressed per pool, whereas the pool had more than one unit.
DR. BRECHER: Right.
DR. NELSON: Question?
DR. SCHMIDT: You touched on some things which relate to old experience, and that is when the bacterial contamination of whole blood was a problem in bottles with rubber stoppers, and all, one of the things that would happen and that you were warned against--I think Dr. Margaret Pitman, at NIH, demonstrated especially with Pseudomonas, was that you if get enough bugs into the unit of whole blood it would use up all the glucose and it would be sterile but tremendously pyrogenic. We see some things here about glucose going down by itself and then going down with the bacteria so that culture would be useless. I guess it has to be at least considered, even though it is probably not happening.
DR. BRECHER: Well, with platelets we don't see that probably because of the short shelf life of platelets. So, even though the glucose goes down you can still culture out the bacteria so they haven't died out, at least in the shorter setting.
DR. NELSON: Yes, go ahead.
DR. KUEHNERT: Matt Kuehnert, CDC. I just had a couple of questions. One was about the slide you showed on BacT/Alert as far as growth. I agree, our data and the SHOT data and other data showed that most virulent organisms grow fairly quickly, but you had one there that I don't think of as a skin commensal and that is Strep. viridans I think. I wonder if you have any comments on why that took so long to grow and whether it was some sort of exception.
DR. BRECHER: It was still less than 24 hours on our inoculation studies. We had five different bottles we were looking at so some of the bottles spiked high but the standard aerobic bottles were less than 24 hours.
DR. KUEHNERT: The other question I had quickly was about notification where a unit had already been transfused. What sort of workup do you do? I mean, in a clinical micro. laboratory you are going to Gram stain it; you are going to go to antimicrobial susceptibilities even. You may not want to do that for P. acnes but for other organisms you would. A clinician might well even request susceptibilities on P. acnes if you report it as a positive. I just wonder how you handle that or plan to handle a situation like that.
DR. BRECHER: Well, we have been notifying our clinicians if a unit went out and we subsequently detected bacterial. We have had four examples of that with P. acnes. In each case there were no clinical sequelae from the P. acnes, which I don't think is much of a surprise. There are only three reports, usually fevers, that have been reported with P. acnes in the world's literature that I can find.
Having said that, there are patients who develop endocarditis and joint infections, eye infections with P. acnes so it is not a completely innocuous bug but usually we don't see problems from it.
DR. KUEHNERT: The follow-up to that, you used an aerobic and anaerobic bottle? Is that right?
DR. BRECHER: That is right. We chose to use two bottles because we thought, and our data suggests, that two bottles are better than one at picking up low concentrations of bacteria. I think we don't have the complete answer to what is the story with anaerobic organisms. It turns out that almost all of the aerobes of interest grow quite well in the anaerobic bottle, with the notable exception of Pseudomonas and Bacillus. So, we have made the conscious decision to use two bottles and that one of them would be anaerobic, at least for the time being.
DR. KUEHNERT: Thanks.
DR. ALLEN: Thank you for you presentation; a lot of useful data. Three quick questions. I wanted to follow-up, first of all, on the potential for the reaction to endotoxin shock versus actual clinical infection that occurs. Does that happen in your experience with platelet transfusions, endotoxin shock?
DR. BRECHER: Yes, it does. With the Enterobacteriaceae often the bottom will fall out with these patients. It is usually the gram positives, that sometime later they become sick.
DR. ALLEN: Any idea of the relative proportion of the two?
DR. BRECHER: Well, roughly 60 percent of fatalities occur with gram negatives, the majority of which are Enterobacteriaceae so most of the fatalities are with organisms that make endotoxin.
DR. ALLEN: Second, are you aware of whether or not the Joint Commission has begun to address this issue in terms of their standards at all?
DR. BRECHER: I don't know that they have specifically addressed it, however, if you are a CAP-accredited laboratory the Joint Commission usually doesn't look in this area very much. So, since the CAP has moved in this direction, essentially de facto JCHO has.
DR. ALLEN: Third, you know, I am not a blood banker and I had sort of made the assumption that most frequently the place of culture would be from the transfusion service rather than the blood collector, although certainly in a hospital you can have both functions together. Is that correct? You certainly implied at one point that in a very complicated, multi-draw system perhaps the place of culture actually might be in the blood collection center.
DR. BRECHER: I think the most impact you are going to have is if you do your culture early, which means the blood center. If we do something late it will have to be a very rapid test. The trade-off is if you have a very rapid test it is not very sensitive, like the dip-sticks and the Gram stains. In the ideal world I think that it should be that whoever is collecting the product, which in some cases is the hospital or the blood center, do a test at 24, 48 hours. If that is negative after a prescribed time period, be it 24 hours or 48 hours if it is a culture system, then I think we can make the case to consider going out to extending shelf life if that is really predictive that a later culture will also be negative, that a negative culture predicts a negative culture.
DR. ALLEN: Given that kind of a system however, does that preclude the pool of platelets being transfused or used earlier if it is needed?
DR. BRECHER: There are actually two paradigms for that. One is, as was described from the Netherlands, you go ahead and let those platelets go out into the inventory, into the world, and only if it is culture positive you call them back, which is the system we have been using at UNC as well as at Dartmouth. The other system people are talking about doing is sitting on those bags for 24 hours, 24-30 hours, waiting until they at least have some preliminary culture data and if that is negative, then they will go ahead and release the units. I think it is going to play out both ways in this country.
DR. ALLEN: We really are talking about increasing the level of complexity of releasing the product however, aren't we?
DR. BRECHER: Withdrawing the units is not going to be easy, there is no getting around that. Yes, it will be more complex.
DR. LAAL: You mentioned briefly something about detection of ribosomal RNA.
DR. BRECHER: Yes, we worked with GenProbe several years ago using a universal probe for ribosomal RNA, over ten years ago. However, that project was shelved. It may be coming back out. There are a variety of other high tech companies looking at various amplification techniques but they are not here yet. There are problems when you are looking at doing an amplification test because many of the enzymes that are used for nucleic acid amplification come from bacteria and, thereby, they are contaminated by some RNA and some DNA from the bacteria. In addition, normal blood running around in those often has some fragments of bacterial DNA. So, it is not an easy a solution as you might think; there are some problems.
DR. LAAL: Still, are there any tests which are looking specifically at bacterial products like endotoxin detection, or something?
DR. BRECHER: Actually, it is funny you should mention that. People have looked at endotoxin in the past. There hasn't been much attention paid to that recently. But I was talking to some of the guys from the FDA and the CDC that maybe we should dust that off and look at it again. It might particularly be more relevant for Third World countries if they can't use this high tech stuff, although dip-sticks are pretty low tech.
DR. KLEIN: Mark, as I understand it you did your cultures in biologic cabinets and hoods.
DR. BRECHER: We did them in laminar flow hoods, which is actually recommended by the company. Some people have done them out on the bench. Jim AuBuchon does his on the bench and his contamination rate is higher than ours.
DR. KLEIN: I think that is the experience in many of the European countries as well which don't culture them in biologic cabinets.
You mentioned cost effectiveness, and I know that cost effectiveness is really only relevant if you extend to seven days. I asked Jim about five days and, certainly, if you retain the platelets for 24 hours after culture there is no way that that is going to be cost effective. I am not sure that is a big point but I just wanted to make sure that is clear.
Finally, in your review of the literature, Mark, and your own experience what percentage of the deaths related to bacterial contamination would be interdicted at 48 hours?
DR. BRECHER: Well, most of the deaths are from the Enterobacteriaceae and we know that the Enterobacteriaceae grow very quickly in these units. So, I think we would probably interdict the majority of the deaths. That would be my guess.
DR. KUEHNERT: Can I just add from our article, it basically would have prevented almost all of them except one. There was a day four, from my recollection, resulting in a fatality. Every other one was either day two or day three storage.
DR. KLEIN: So, that is culturing on day one and releasing--
DR. KUEHNERT: I am sorry, I thought the question was about transfusion fatalities from bacterial contamination. So, day two or day three of storage was where almost all the fatalities were.
DR. BRECHER: And that is similar to the SHOT data which implies, as I said, that when you look at growth curves these bacteria reach high levels within two or three days. Fortunately, they are also picked up very quickly with these culture systems and they were all gram negatives. That is another important point.
DR. NELSON: Mike?
DR. BUSCH: Just to update people, Blood Systems or UBS is planning to have bacterial screening in place by July 1st, including apheresis and randoms. Just to address the problems we are facing with the randoms, just as you indicate, right now you need to culture each random separately and you need to leukoreduce them before you can culture them. So, the cost to culture separately four, five or six randoms will actually drive the price of pooled randoms well above the price of apheresis. That is one of the issues.
The other, in terms of your cost effectiveness, as Harvey was commenting, right now we run outdates of 5-12 percent in different regions. The problem is going to be that we are going to have a mix of BacT/Alert and Pall going on but with the BacT/Alert systems we are going to be inoculating those cultures in the component labs in the regions and then sending the culture bottles to the central lab that will do the culturing. So, the inoculation won't happen until 24-30 hours after collection. Then you have the transport. That transport of the bottle doesn't count in the 24-hour culture and we plan to only label after these cultures have completed 24 hours. In essence, platelets are not going to become available for labeling and issue until approximately three days. So, I don't think we are going to see any savings. I think it is going to be the opposite until we get an extension of shelf life outdates and the potential crisis of platelet availability is going to be the challenge.
DR. BRECHER: Right, and I think there has been no incentive to look at platelet solutions because the shelf life is limited to five days. If we open the door we may be able to figure out better storage solutions for platelets. Maybe we can eventually get to ten days.
DR. NELSON: However, you said that the primary reason for going from seven back to five was the bacterial contamination issue. So, if that is solved--
DR. BRECHER: Very simplistically, that is correct.
DR. KLEIN: That was in apheresis platelets. There aren't really any data on seven-day pooled whole blood-derived platelets.
DR. NELSON: Comment? Please state your name.
DR. LEE: John Lee, FDA. Given your comments about the growth kinetics and that the plateau phase might be reached in the first two days with most bacteria, how do you reconcile that observation with the fact that a flurry of additional bacterial contamination related fatalities occurred when the shelf life was extended from five to seven days?
DR. BRECHER: Well, I think that that was very anecdotal. There are only a handful of cases that were reported, mostly from Johns Hopkins back then. I think that it may have been that people were looking a little more carefully because the platelets had just gone out to seven days and people were paying a little more attention, but I really don't have an answer for that.
DR. NELSON: Thanks. Yes?
DR. STRONG: Just another comment on the logistical difficulties of this. Mike points out that we will have a delay in release which will increase the outdate and perhaps reduce the availability of apheresis platelets. I have trouble with that term "random" because actually they are all random. As you point out, we don't have a good culture remedy yet because we can't pool up front and we can't store for seven days. Clearly, the standard that has now been set requiring testing of all platelets may result in a shift towards more apheresis platelets. You might argue that that is a good thing because they are safer. On the other hand, we won't have enough to begin with.
We have heard from our hospitals that if they are going to be forced to test, such as dip-sticks or Gram stains, or whatever, they are not going to do that. They are going to order apheresis platelets, which we don't have enough of. So, there is a potential downside of all of this in that we will have a platelet shortage crisis. Mike has already brought that up. So, unless we can get seven-day platelets I think we have a potential real problem on our hands.
DR. BRECHER: Right. I think we all need to step back and ask what is best for the patients, what is best for the country and what does the data support? I think there is a lot of data that we have heard today that suggests that we can probably pre-pool and we would have the safest platelets. So, we would have done everybody a favor I think if we can move in that direction.
DR. NELSON: We are sort of fighting the lunch deadline but Dr. Katz wanted to make a statement and Kay Gregory, and then we will break for lunch and we will continue the discussion later.
Open Public Hearing
DR. KATZ: I want to point out a typo in our printed statement. The first line of the second paragraph says 1970s. It is supposed to say 1980s.
I am representing America's Blood Centers, a national network of locally-controlled, not-for-profit community blood centers that provide nearly half of the U.S. blood supply from volunteer donors. Collectively, ABC total blood collections exceeded seven million donations in 2001. We operate in 45 states and in Quebec and serve more than half of the 6000 hospitals in the United States.
In the 1980s the platelet outdate period was extended from five to seven days, and subsequently rolled back to five days when it became apparent that an increase in septic reactions to platelets was associated with the longer outdate. This was appropriate.
Today ABC members are pleased to participate in the efforts of the blood community to control septic reactions to platelets using currently approved systems for bacterial detection. It is in this historical context that BPAC and FDA should address both extension of platelet outdates and the regulatory approach to allowing pre-pooling of platelets from whole blood, that is random donor platelets. The deployment of these detection methods should allow reconsideration of longer platelet outdating.
ABC endorses the comments that you will hear from AABB primarily addressing pre-pooling. We further emphasize the importance of expeditious and non-burdensome attention to extension of platelet outdating to seven days and beyond potentially as a measure to enhance safety and availability of platelets to patients in need. The optimal use of bacterial detection systems requires a one- to two-day incubation using the licensed quality control systems available, and incubation cannot begin until 24 hours after component preparation. This will leave only one to two days of shelf life on a five-day platelet product, and may impair the availability of these products most critically for blood systems and transfusion services dependent on imports from remote collection facilities.
The FDA should use the least burdensome approach to allowing extension of outdating for those products subjected to sensitive bacterial detection. We support the need for limited clinical studies of platelet recovery and survival to replicate the data that led to a seven-day expiration in the 1980s using current collection methods. For apheresis platelets, autologous studies will be acceptable. For pooled random donor platelets, studies in thrombocytopenic patients will be required. They should be statistically powered to demonstrate pre-agreed upon recovery and survival in comparison to one-day old platelets with pre-agreed statistical confidence, and not be unnecessarily large so as to inhibit the adoption of bacterial detection systems.
Clinical trials using prevention of bleeding as an endpoint are not needed since recovery and survival are broadly predictive. In vitro studies of platelet function should not be required since their correlation with clinical endpoints is weak, and Dr. Snyder has already shown you that they are reasonable.
The voluntary implementation of bacterial detection by the blood community represents a robust approach to the most serious current infectious risk of transfusion. FDA's participation as outlined above is in the best interest of the safety and adequacy of the blood supply. Thank you.
DR. NELSON: Thank you. Kay Gregory, from AABB?
MS. GREGORY: The American Association of Blood Banks is a professional society for over 8,000 individuals involved in blood banking and transfusion medicine, and represents approximately 2,000 institutional members, including blood collection centers, hospital-based blood banks and transfusion services as they collect, process, distribute and transfuse blood and blood components and hematopoietic stem cells. Our members are responsible for virtually all of the blood collected and more than 80 percent of the blood transfused in this country. For over 50 years, the AABB's highest priority has been to maintain and enhance the safety and availability of the nation's blood supply.
The AABB strongly encourages the FDA to allow pre-storage pooling of platelets derived from whole blood, and further encourages the FDA to request only scientifically and medically reasonable data to support such a change in approved practice.
The need for pre-storage pooling arises from the desire of blood collecting facilities to perform cultures of all platelet units to interdict those that are bacterially contaminated. Blood bankers recognize that bacterial contamination of platelets is the greatest residual infectious disease threat to transfusion recipients and causes both frequent deaths and substantial morbidity. Improvement in skin cleansing practices and diversion of the first few milliliters of collected blood are being implemented to reduce contamination of components by skin flora.
Currently, however, the most effective way of dealing with this significant problem is to perform a bacterial culture on each unit. This is already feasible with apheresis platelet units, but culturing individual platelet units derived from whole blood imposes serious additional difficulties. The volume needed for the culture to capture contaminating bacteria will remove a significant portion of the small volume in a unit of platelets from whole blood, thereby reducing its clinical efficacy. Furthermore, the sheer number of cultures to be performed will overwhelm the technical capabilities and financial resources of the transfusion system. Blood bankers want to address this issue but we need a practical way of doing so.
The AABB believes there are already adequate data to support the practice of storing pooled platelet concentrates for the remainder of the usual storage period of its individual units, providing that bacteria are reliably detected in the pool. First, ample data are available and have been summarized today by Dr. Snyder that this pooling does not lead to immunologic activation between donors such that a mixed lymphocyte reaction occurs.
Second, the sterile connections through which the pooling will be conducted have already been discussed by this committee. BPAC determined that appropriate use of a sterile connecting device will not increase the likelihood of contamination. The resulting pool of platelets will need to be stored in a larger bag, of course, to facilitate the necessary gas exchange during storage. Manufacturers of bags used for storage of apheresis platelet units already have documented the capabilities of their bags to hold the requisite volume and platelet content of apheresis platelet units, and we believe one can reasonably extrapolate these data to whole blood-derived platelets.
Finally, we must not ignore the experience of others. Blood centers across Western Europe have been pooling platelets derived from whole blood for many years very successfully. Although these are usually derived from buffy coats, they provide an adequate, safe system for platelet storage. There should be no need to conduct studies to document again the in vivo recovery and survival nor the clinical utility of using pooled platelets.
The agency is concerned that the increased volume of a pool of platelets, in contrast to a single unit's volume, would allow contaminating bacteria to generate a larger inoculum and, thus, represent a greater risk for patients. The scientific basis underlying this concern is not strong, particularly when the endpoint concentrations are considered. The increased volume of a pool would make a difference only when the bacterial growth had reached a limiting concentration, and this concentration is usually in the tens of millions of bacteria per milliliter and higher. The difference in the total inoculum of bacteria would be higher than that from a single unit, in direct proportion to the number of units in the pool, according to the FDA's concern. However, at such high concentrations the morbidity is likely to be dramatic regardless of whether the pooling was conducted before or after storage.
The AABB recognizes that authorization of pre-storage pooling will require a reliable system to detect bacterial contamination. The systems already approved for in-process quality control testing offer this capability, and others are known to be under development.
Although bacterial contamination is a very significant threat to platelet transfusion recipients, its relative rarity makes any study conducted on platelets from whole blood platelets collected and stored in the routine manner extremely difficult to study. The AABB suggests that the in vitro or spiking studies that have already been conducted be referenced as the primary basis of acceptance of a bacterial detection system that could be used to allow pre-storage pooling. This would be analogous to using not blood donors but a population at high risk for HIV to document that a new HIV antibody test was sensitive for detecting that pathogen's presence.
The FDA has not based its approval of HIV test kits on the documentation that when donors were tested a second time they remained negative. The effectiveness of these systems could be tested in the real world by retesting units at outdate after the initial culture to determine whether the initial culture was accurate. However, the enormous size of such a study, suggested to involve reculturing 50,000 to 100,000 units, will effectively preclude the study's performance. The result will be that a significant proportion of platelets transfused in this country will never be cultured for bacteria, and an opportunity to enhance patient safety will be missed.
Instead, we urge the agency to look at the entire problem, that of bacterial contamination as it currently exists and as it could be markedly reduced by culturing pooled units, and exercise the FDA's mandate to improve the safety of transfusion by taking reasonable measures.
We would also add that the ability to leukoreduce multiple units of platelets through a single filter as part of the pooling process will provide an economical way of conducting pre-storage leukoreduction with all the advantages this committee has already recognized. Since the vast majority of platelet recipients benefit from leukoreduction, and since it is generally accepted that removal of leukocytes prior to storage minimizes the likelihood of febrile reactions, allowing filtration and pooling would be a boon to recipients even beyond the facilitation of culturing.
Finally, the addition of bacterial detection systems will necessitate some delay in release of platelets, which may ultimately result in a higher frequency of outdates and wastage. Given that platelets are already in short supply in some areas of the country, a further decrement will worsen these shortages unless the storage period is increased to at least seven days, where it was in the '80s before recognition of the importance of bacterial contamination was recognized and we went back to five-day storage. Extension of the dating period for bacterially tested pooled platelets should be encouraged. Thank you.
DR. NELSON: Thank you, Kay. I think I would like to break for lunch now and come back at 1:30 and continue the discussion. Thanks.
[Whereupon, at 12:35 p.m., the proceedings were recessed, to resume at 1:35 p.m.]
A F T E R N O O N P R O C E E D I N G S
DR. NELSON: Nobody else had asked to speak in the open public hearing but if anybody has any comments--if not, I think we will proceed to the questions for the committee. Dr. Vostal?
Questions for the Committee
DR. VOSTAL: Thank you.
We heard a lot of data this morning concerning how to evaluate platelets and how to evaluate the bacterial detection devices. We are hoping to get some input from the committee on these two topics.
One of the discussion points we have is that we would like to ask you if you could please comment on the FDA proposed plan for validating an extension of the storage time for pooled platelets. In general, we would like to have input on the assessment of platelet quality and validation of the bacterial detection system.
DR. LEW: It was brought up earlier that at one point FDA had approved seven days. I guess I would just like to know has processing changed dramatically since then, that now it is really worthwhile having to repeat everything? Or, could there be just limited studies that you are recommending based on the fact that it has been approved before?
DR. VOSTAL: Processing has changed a lot, especially for the bags that are used to store platelets and the gas exchange of those bags. So, since the differences are significant, we would actually like to see studies that demonstrate that the current storage conditions are adequate to support platelets out to seven days.
DR. FALLAT: We heard a lot of data about the five- to seven-day storage done currently, but I guess not within the framework of the FDA. How much of that data--is that enough support for the in vitro? Not being in this field, I can't really answer that but I certainly saw a lot of good data, it seemed to me.
DR. VOSTAL: Currently in the U.S. we only have five-date dating for platelets. In Europe there are countries that store their platelets out to seven days. However, as you saw from Dr. Pietersz, the collection of those platelets is different and the platelet product, called the buffy coat platelet, actually is different. So, one of the issues we are trying to figure out is whether that data for those platelets would apply to the U.S. situation. If it does apply, it would make it a lot easier for us because of their clinical experience.
DR. NELSON: I wonder if you could summarize--I am not clear on what the FDA proposed plan for validating these three issues is. Could you summarize that? In other words, what are we supposed to comment on?
DR. VOSTAL: Yes, sorry. For evaluating platelets we have a standard procedure that we follow, and we have a platelet testing guidance. Usually it involves in vitro testing on the platelets at the end of their storage period. Then, we go to radiolabeling studies. However, in this case it is difficult to do those studies because you have a pooled product and you would be giving that to a single volunteer donor and there are ethical issues involved in that.
To get around that, we are proposing that studies to evaluate in vivo efficacy of these platelets would take place in thrombocytopenic patients who would get transfused with a pool and you would follow the outcome of those transfusions by count increments and corrected count increments. So, for validating the storage bags themselves, that is our approach.
DR. NELSON: And what is the N in that trial, or how many people do you think would have to be studied to validate the bags?
DR. VOSTAL: Those studies will probably require somewhere around 50 patients. There have been studies done, for example, there was a study done by Baxter in Europe that looked at the efficacy of psoralin-treated platelets. They did a study where they followed CCIs. That was a study that had a total number of 100 patients, 50 per arm.
DR. NELSON: Clinical efficacy I guess would require a larger number, right?
DR. VOSTAL: This is one of the difficult questions. I mean, the clinical efficacy of platelets is really their ability to stop bleeding or their ability to prevent bleeding. For changes that we consider relatively minor stress on the platelets, we take the ability of platelets to circulate as a surrogate to efficacy. But in situations where we feel that the damage could be extensive, for example, if you are introducing a pathogen reduction treatment that may be causing all sorts of damage to a platelet, then we think that is enough of a concern to go to a study that follows hemostasis and bleeding.
But in this area, I think based on the European experience with pooled platelets and no problems with hemostasis in Europe, I think the concept is demonstrated, such that we probably will not ask for studies to look at hemostasis.
DR. NELSON: Count increments would work?
DR. VOSTAL: Right.
DR. DAVIS: It sounds to me like you want to reinvent the wheel, but how long would it take to revalidate the information that you are asking for here?
DR. VOSTAL: Well, I would guess that if you started today you would probably, you know, do in vitro studies and then set up a clinical trial, I think probably within 12 months you could get an answer.
DR. LEW: Have there been any studies already just comparing the European way of collection with the buffy coat and compared to how it is done here, to kind of help guide or give us a hint?
DR. VOSTAL: I am actually not aware of any studies that compared platelets prepared by the PRP method and prepared by the buffy coat method. Dr. Snyder, are you aware or any?
DR. SNYDER: There were a fair number of studies that compared them in vitro. I think the point you asked about is in vivo and transfusions, and I don't think there were a lot of those. I would have to go back and look, but maybe Dr. Pietersz would be able to comment on that. I think most of those were looking at in vitro characteristics. Do you agree?
DR. PIETERSZ: I know that Stan Holme did a lot of comparisons between PRP and buffy coat platelets, but they were in vitro studies.
DR. VOSTAL: I think the big difference between the PRP platelets and the buffy coat platelets comes when the PRP platelets are actually centrifuged hard in the bags so there is a platelet pellet which is then resuspended. The buffy coat platelets are always centrifuged on top of a red cell cushion. I mean, this is just my opinion but I think the buffy coat platelets are less stressed during the collection procedure.
DR. KLEIN: I think that is right. There is a lot of in vitro data that shows that there may be a slight advantage to the buffy coat system. There are no head-to-head prospective transfusion studies.
I just want to comment that I think we have seen a load of in vitro data on pooled platelets stored for five days or seven days. We know that in vitro data in general is not predictive of performance in vivo. Personally, I can't see any reason to do additional validation studies for pooled platelets in vitro.
On the other hand, there haven't been studies of pooled platelets, seven days, in vivo and I quite agree with you, based on the European data I think we can feel confident that a relatively small study, looking at corrected count increments, could tell us exactly what we need to know to go ahead with these kinds of applications.
DR. NELSON: I think the FDA would have to perhaps set some guidelines. In other words, I am not sure that if there was a statistically significant difference in the corrected counts that that would be clinically relevant because you could give a larger volume of platelets and make up for that. You know, maybe the FDA has this guideline but I am not aware of it, and they need to set some guidelines as to what is an acceptable comparison between the five- and seven-day storage.
DR. KLEIN: I would be astonished if the seven-day platelets were exactly as good as the five-day platelets, and certainly they are never going to be as good as two-day platelets. I think everybody accepts that. But I guess the real issue is whether they are safe and effective. I am not going to design the study for anyone, even if a company supports it, but I think that you can very easily say five percent, ten percent, whatever you want, to say less than a five-day old platelet would be effective. If you can define effective, they will be effective; they will stop bleeding.
DR. NELSON: We heard the numbers 75 percent and 80 percent and 70 percent; you know, you pick a number I guess.
DR. ALLEN: We have heard a lot of data and, yet, most of it is data that is several years old to five to six years old, looking at some of the containers, devices and the filters that are currently available. The studies are always running well behind available and approved capabilities. We heard Mike Busch comment from the floor about Blood Systems proposed introduction of bacterial culturing of pooled platelets and that that would be done through a centralized laboratory.
The processes are changing very rapidly. So, I think somehow the FDA has to get in hand the information that it needs to make some of the decisions, and needs to encourage broad-based studies. Again, I am sorry that the full committee isn't here, but this may be an area where either the FDA should be funding some very targeted studies or NHLBI ought to be doing so, or other sources of funding but I think the FDA needs to work, as it has occasionally in the past, closely with academic institutions to design and implement the studies. When I say academic institutions I am certainly including the large blood collectors that have done this kind of research. I think we need to evaluate and collect data.
I also think we need some way of realistically looking at the trade-offs such as Dr. Klein's statement that seven-day platelets, while they may be on a simple scale both safe and effective, probably aren't as good as five-day old, and those aren't as good as two-day old. I mean, from my perspective, although I haven't done clinical medicine in a long time, I would much rather get one- or two-day old platelets than five-day old platelets if I had to be a platelet recipient.
We are talking about mechanisms that in fact are going to push that availability and use on out. So, I think we have to look at what are the trade-offs there. Yes, maybe it can be approved from a regulatory perspective but, as clinicians, we would always like to see them used or be made available at an earlier point. So, I think we have to be cautious about making final decisions about how long a culture has to be run before it can be released, and that sort of thing, and I think those are regulatory issues that the FDA has to wrestle with and I am not sure that we have nearly enough data today for us to give you good advice nor for you to make a really meaningful decision.
DR. NELSON: On the third point, the validation of a bacterial detection system, was this referring to the BacT/Alert or culture based system, or some other system?
DR. VOSTAL: Right now the only systems available are culture based, the two that are on the market. So, currently our studies are sort of geared towards validating those systems. If a system comes up on the market that has almost real-time recognition, real-time readout of the bacterial contamination I think studies could be designed a little bit differently.
DR. NELSON: Was the FDA prepared to accept data, either partial data or a whole study, where the bag was intentionally contaminated with a certain number of level of organisms that are commonly found, or were you talking about essentially a natural history study? What sort of validation were you aiming for here?
DR. VOSTAL: Well, what we have proposed so far is to do laboratory studies where you actually spike units with small amounts of bacteria. But we also feel that besides that we would like to see a field study that samples naturally contaminated units and you would get an idea of how sensitive that device is in the real-life situation.
DR. NELSON: I guess one of the things that a natural field study would tell you that a spiked study wouldn't would be the frequency of false positives. You can learn that but the true positives might require a fairly large study I guess if you were looking at the range of organisms, etc. I think that is a constraint. But I would think that, in addition to a spiked study, at least a study that would allow some estimate of what the false-positive rate is, given a new method for the pooling and the extended shelf life, etc., and a new bacterial detection system might be worthwhile to learn what the false-positive rate is because that would influence decision-making, wastage of platelets, etc. and might uncover some problems.
MR. VOSTAL: That would be a lot smaller study because you would only reculture the positives that you would get. So, the question is would that be sufficient, or should we wait to see results of a large field trial?
DR. DAVIS: I would like to say again that as a blood user I have seen enough to convince me that I think it is safe enough to go ahead we pre-pooled platelets and keep them in storage for seven days. If it increases the supply of platelets, I am all for it. I have been in situations where I have been out of blood; I have been out of platelets and out of components. I don't want to be in that situation again.
DR. KLEIN: I want to get back to the point that Jim Allen made a little earlier, that is, the issue is not whether the seven-day platelet is not as good as the two-day platelet but, really, I think what we are faced with is whether a seven-day pool of bacterial cultured platelets is better overall in terms of safety and efficacy than is a five-day pool at the time of issue that isn't going to be tested--those are not going to be tested for bacteria. In fact, what will happen is a mandate for bacterial culture or testing, they will disappear and you will only have apheresis platelets. I think that is the practicality, which several people can comment on who are involved in large collection centers. That would be a major issue in terms of platelet supply in the United States because somewhere between 25-30 percent of all platelets are today whole blood-derived, so-called random units.
DR. STRONG: Just to amplify, the last two speakers have spoken on the issue of supply and I think that is critical. We shouldn't be striving for the perfect in this situation. We really are on a timeline here. We have a standard that has been set that needs to be implemented by March of 04 to do testing on all platelets. There isn't a system available right now for the whole blood platelet, which will result in forcing us to apheresis platelets. We really need to find a way to allow us to continue to use whole blood platelets in a way so we won't jeopardize the platelet supply.
There are a couple of issues that I think need to be addressed. One is that the FDA hasn't set a good standard for what needs to be achieved in any kind of study that we do. Should it be 75 percent recover, 50 percent survival, a CCI of X, whatever it may be? That is the place to start so that the studies can be begun. But if we can have a process that would allow us to implement a pooled platelet product that we can bacterially test while we collect the data, and actually I think there is a lot of data to suggest that it would be okay, but allow us to collect scientifically valid data in the interim period so that we don't have to wait for the three to five years that it would take to collect scientifically acceptable data to allow this to happen because that will be too late.
DR. GOLDSMITH: Could you comment on the impact of the timing of the leukoreduction on the proposed trial designs as an additional variable?
DR. VOSTAL: Well, I would agree with Dr. Snyder that pre-storage leukoreduction is optimal and the longer you wait, I think the less benefit you would get from leukoreduction.
DR. NELSON: That is what has been recommended I think by BPAC. I don't think it is mandated but it is recommended at the present. Yes?
DR. FITZPATRICK: A couple of things. I think we need to be clear that there are two purposes for platelets. One is prophylactic prevention of bleeding in a patient. The other is to stop bleeding in a patient. Counts and CCIs don't really address efficacy in both those situations.
I think Dr. Snyder presented us with a bulk of data on different days of storage, even out to eight and nine days. So, in vitro data shows that there really isn't an in vitro difference in different bags, in different situations, in different collection systems. In systems that are currently used from an in vitro perspective, the platelets look pretty much the same. There is a bulk of data on that that tells you that in vitro there isn't much difference.
Clinically and effectively, physicians I think will tell us that there is a difference in some patients as the platelets age. The older platelet may not be the best platelet for a frankly bleeding patient but we don't know that for sure and we don't have data to support that, or good studies to support that.
The question on validation confuses me a little bit because there are two FDA approved devices for detection of bacteria in platelets that are approved. So, one type of validation is user validation, is the user using that detection system in the way the manufacturer intended for getting the appropriate results? I don't think that is what you are asking us. I think what you are asking is what should the FDA do for the user to validate that he can use that system and keep those platelets for seven days.
In your presentation you suggested filtering at 24 hours or time of expiration at seven days, and doing a whole lot of units to do that. You heard from organizations saying that they thought that might be burdensome and increase the costs. I would agree with that. It would be burdensome to do that many units and increase that cost and you should maybe look at some spike assays for that with slow growing organisms, but if it is negative at 24 hours and you have evidence, with some statistic sampling, that it is negative at seven days we shouldn't ask every user to do that for 10,000 units.
So, I think you are asking two types of things about user validation. If it is the user who is using it as the manufacturer intended, then the validation should be relatively simple.
The other aspect of that, which I bring up somewhat tongue in cheek, is the fact that there is an application before FDA for a frozen platelet that can be kept for a year. It is my hope that we will be able to get that approved within the next few months but it requires a multi-center trial which would require funding. I would hope that we will be able to meet with the FDA soon to do that.
But if there were available to the clinical population a platelet that could be frozen and easily thawed and used within minutes, not hours, for the frankly bleeding patient that would help with the supply problem. It has been shown to be effective in one clinical study, not in multiple clinical studies, and is an aspect where we should possibly start thinking about two different types of platelets for use in patients and alleviate some of the strain on the liquid platelet inventory by progressing as quickly as possible by getting this system licensed.
It is my understanding right now that the only proposed regulatory body setting the standard--we have CAP and AABB--but the FDA has not proposed a guidance or a rule on bacterial detection in platelets. So, even within those standards we have the risk of a tiered system, as was talked about by the presenters. We have swirling; we have dip-sticks; we have multiple different methods of detecting bacterial contamination of platelets, some much more sensitive than others but some a lot less expensive than others.
So, without rule-making or a guidance, and with the pressure on supply and inventory, and without the pre-pooling you will essentially mandate a two-tiered system in order to keep supply available. I don't think that apheresis platelets will just automatically replace random donor platelets because of cost pressures and because of supply pressures. So, because they have to have a system someone will use swirling, or someone will use dip-sticks. We know that those systems, while they may be better than nothing, are not the best available.
You know, we have ample evidence of in vitro stability of the product. We talked last time about the sterility of the sterile connection device. Perhaps we could move forward with pooling at least for five days and collect information on seven days. Maybe there is a way with postmarket information to be able to do that, or some focused studies at sites with seven days within NIH. But I think pre-pooling offers great advantages both to the patient and the center, and might help alleviate blood type shortages if the collection centers pool in pools of like type and then we don't have problems with the hospitals just picking whatever is on the shelf and pooling it. I think there are a lot of advantages to that.
DR. VOSTAL: You made several points here--
In terms of the validation issue, what we would like to see is validation done by the manufacturer to support their intended use. So, we would like to see data. Currently, the two devices are validated for quality control of platelets.
Our thinking is that if they would like to have a platelet screening or release of platelets as an intended use, then they would have to go through this field trial. But I am not talking about validation done by the actual user; this is for the manufacturers.
Then you talked about different kinds of platelets, having platelets for prophylaxis and having platelets for treatment. I think there could be two different indications for platelet products. However, I think currently platelets that we transfuse basically are used for prophylaxis use, to be able to circulate so if there is a breach in the cardiovascular system they will be able to plug that up.
You were talking about frozen platelets, I think we recognize the need for that for treating bleeding soldiers in the field and for acute treatment use. It might be something that may not be a platelet but could be a platelet-like product or something like that.
In terms of being able to put together a guidance on bacterial detection in platelets, I think it is in the works. We are trying to put one together and, hopefully, we will get one together relatively quickly.
DR. KLEIN: I would point out that we do have a standard really, a generally accepted standard for platelet transfusion. Whether it is by increment or corrected count increment, every clinician uses the generally accepted standard and I think that wouldn't be difficult to agree upon.
I think the difference is whether this particular platelet stored at seven days is as effective in terms of increment as two-day platelets. I think as long as we don't have that mind set, as long as we look at the entire picture and say is it safe or is it bacterially tested and stored, and if it isn't and still meets a clinically accepted standard--patients are quite different than normals who receive labeled platelets--I think we can get a pooled product in an in vivo trial with an acceptable endpoint and demonstrate that these are good, as I suspect they are from the European data which isn't exactly the same, or whether they are not. But we do have a standard to look at.
DR. STRONG: I would also like to comment on your encouraging the manufacturers to do these studies because it has been our experience, at least thus far, that there is really no incentive for them to do that. They have de facto got the market to do what we are doing because the blood centers have decided to do that. So, there really is no incentive for them to spend the money to do all of the kinds of validations that you are talking about. So, somehow we have to get past that as well because the centers themselves cannot afford to do it.
DR. SNYDER: Two comments, one relating to frozen platelets. Frozen platelets are something that I think is a good goal, however, it is logistically difficult for hospitals to deal with that. I assume it would be in DSMO they would be frozen, although I don't know for sure. If it is DSMO, that would be a problem. We would need IRB approval for sure because most noon-cancer patients would not necessarily want to get DSMO without knowing about it and the time to ask them about it is not when they are in extremis, etc. There would be some logistic issues, though it certainly would be useful. I wouldn't want to delay studies looking at what we can do better with liquid storage while we push a frozen protocol, not that you suggested that but I just wanted to say that.
The other thing related to looking at five-day and then going to other studies to look at seven-day, if you can sit down with a large enough cup of coffee or chamomile tea, depending on your preference, you probably could figure out a protocol where you could pool--and that is something else the agency needs to give some guidance on, how many units are you allowed to put into a pool--if you could pool, let's say, eight units together and then you could divided it into two aliquots, one aliquot could be transfused into the so-called standard sick thrombocytopenic individual at day five and then you have all of day five, all of day six and all of day seven. You could then transfuse on day seven, and these patients normally require platelets perhaps three days later, the seven-day stored of the same pool. Then you could, therefore, get five- and seven-day data that could be comparable.
Although that would be difficult and you might limit the number of patients, it might in the long-run be more efficient to consider doing it that way and at least you would get five- and seven-day data. At most academic centers you probably could put together a protocol like that.
So, I think these are some things we should think about. I would hate to just do five-day and then have to come back and start all over again to do seven. We should be clever enough to figure out how to do, you know, baby and bath water at the same time.
DR. NELSON: That is a good suggestion. Other comments? Have we commented sufficiently?
DR. VOSTAL: We actually had three discussion points but I think we have run through most of them. Could I have the next slide?
The second one was to please comment on the applicability of the European experience with pre-storage platelet pooling of buffy coat platelets to the FDA proposed criteria for approval of pre-storage pooled PRP platelets.
DR. STRONG: I would like to encourage us to be able to do both buffy coat and the current PRP approach. There is some data to suggest buffy coat platelets in fact might be better, particularly for extended storage, because of the things that you just talked about. But we should be allowed to explore both of those opportunities.
Clearly, the European data is now coming to light, particularly the bacterial testing as well as clinical efficacy, that would suggest that it works. How we validate in this country is kind of the issue and we need to be able to move forward in doing that.
DR. NELSON: I have one other question with regard to the bacterial contamination. Is it now FDA mandated, the initial diversion and the skin prep, or is that still whatever anybody wants to do? I know the AABB has made some recommendations but that might affect the frequency of bacterial contamination, particularly from skin flora. Is that mandate or recommended by FDA now?
DR. VOSTAL: It is recommended by the FDA but not mandated. Actually, that was an issue we discussed at BPAC maybe a year ago, diversion pouch, and we put forth criteria for approval of diversion pouches and we have received several applications and several are already on the market.
DR. NELSON: So, do you think most blood collection facilities are doing that now?
DR. VOSTAL: Actually, I am not sure how many are involved.
DR. STRONG: It partly relates to the availability of bags so there are still licensure issues.
DR. KLEIN: Getting back to this discussion point, I think you have four issues that you began with, seems like days ago--
--about pre-storage pooling. One of them had to do with the reaction between lymphocytes, the MLC. I think to answer this question, the European data, I feel very comfortable, has set that aside. The second had to do with the quality of the platelets because of possibly antiplatelet antibodies, etc., etc. Once again, I think I feel very confident that the European data is quite acceptable in that regard. The third had to do with bacterial contamination and, since we are going to be testing, we don't have to rely really upon the European data for that. The fourth one, I confess I have forgotten.
DR. VOSTAL: It was the sterile connecting device.
DR. KLEIN: So that is no longer an issue as well.
DR. ALLEN: In the United States, does the FDA now require the PRP platelet preparation and does not approve the buffy coat? Is that correct? I mean, that is not an option to United States blood banks?
DR. VOSTAL: Right, as far as I know, the buffy coat platelets are not licensed in the U.S.
DR. ALLEN: I agree with Harvey Klein's comments. Based on the data we heard today, it certainly seems as though that is a very reasonable option to offer. I am not sure, from a process perspective, what it would take on the part of U.S. blood banks to switch over but I have no reason to think that it should not be available as an option.
DR. SIEGEL: We don't mandate against buffy coat platelets, just that blood centers have not come forward requesting approval of buffy coat platelets.
DR. KLEIN: That would really be a sea change in how blood centers function. That is not a minor change.
DR. WAGNER: Steve Wagner, from the Red Cross. I just wanted to make a comment about buffy coat platelets. There are some differences in practices. To get a really good, high quality buffy coat from which you can make platelets you really need to hold whole blood for an extended period of time. Most of the European countries that are making buffy coat platelets hold the whole blood for more than eight hours, which is what we are currently allowed in the United States to hold whole blood. So, there is a variety of things that would have to change in order to be able to practice buffy coat platelets.
In addition, when you make buffy coat platelets you are going to lose a certain number of red cells. So, that also has to be factored in. So, it is not an idle discussion to talk about conversion, but it would require quite a lot of work and a change in the way blood centers and blood providers operate, and the way in which the regulations now stand, as I understand it.
DR. STRONG: I would agree with those comments. Clearly, it would be a huge process change for all of us. On the other hand, if it ultimately leads to improved platelets and perhaps even longer storage times, and there are data I think in several laboratories to suggest that the buffy coat platelets may be storable for longer periods of time, particularly with additive solutions which somebody commented on this morning, that we should encourage more experimentation with additives to extend the platelet storage time. Nevertheless, we have been through pretty major changes before in blood centers, NAT not being the least of those. So, I think it is entirely doable. It is just a matter of having the ability and the receptivity by the FDA to do it that would perhaps encourage some of that.
DR. VOSTAL: Maybe we can move on to the last discussion point.
In case we didn't cover some of these things, if you could please comment on the following bacterial validation issues, how useful is the bacterial detection data on buffy coat platelets collected in Europe to validation of bacterial detection devices for PRP platelet pools?
If bacterial detection devices are validated for PRP prepared random donor platelet units, do they need to be revalidated for PRP platelet pools?
If bacterial detection devices are validated for leukoreduced platelet products, do they need to be revalidated for non-leukoreduced platelet products?
DR. NELSON: I think with regard to the second question there, you might get more false positives and true positives actually in a pool. But I would be concerned about whether the pooling actually did something to the false-positive rate. Maybe I am wrong but that could certainly influence how platelets were handled.
DR. ALLEN: Again, I go back to the point I made about every time you change the process, even in a small way, it may have an impact. For example, if we find that from a year from now a large proportion of all blood being collected has that initial diversion system in place, perhaps the potential for contamination from skin bacteria and skin plug bacteria may increase significantly.
Given the systems that I have seen demonstrated on slides, they look fairly complex. Nonetheless, they have the sterile connector systems in place. I think they are trying to make them people-proof, which is probably good. Nonetheless, they look reasonably complex to use and my hospital infections experience from many decades ago suggests that the more complex they are, the greater the chance for somebody to screw up somewhere.
So, I am not sure that they need to be totally revalidated. Nonetheless, I think the FDA should encourage studies of each new process change and that ought to be a continual event. Even once you get useful products and systems on the market and in use, we need to continue to look at these.
To my mind, I think there are a number of questions in terms of the bacterial contamination, one being can we reliably detect bacterial contamination very early without waiting for that 24-hour sampling period?
Second, if you are pooling, are you then reducing the amount of the volume that might come from any single contaminated unit and what would that do to the sensitivity of being able to detect, or the rapidity with which a system, such as the radiometric assay, would become positive?
So, I think for each of these stages questions need to be raised and be looked at. I am not sure I would necessarily suggest that they have to go through a whole revalidation process but I think part of the academic inquisitiveness that is the essence of the scientific method on which our practice of medicine should be based suggests that we ought to be continuing to do studies of this type.
DR. NELSON: One question is are the same bacterial detection devices used in Europe? You mentioned there are two licensed, are these the same in the U.S. and Europe?
DR. VOSTAL: Actually, I am not sure which ones are licensed in Europe. The BacT/Alert is being used in Europe.
DR. NELSON: Right.
DR. VOSTAL: I am not sure if the other ones are.
DR. KLEIN: Again, in focusing on the questions that you are asking here, if you remember the slide that Ed Snyder showed comparing the way you prepare platelet-rich plasma and platelets compared to buffy coat, they are not that different. One is a hard spin first and one is a soft spin first. The quality of the platelet may be a little different but basically the component is extremely similar. So, in terms of how useful is the data from Europe for platelet-rich plasma-derived platelet concentrates, I think they are very relevant.
On the other hand, if you are talking about leukoreduced versus non-leukoreduced platelets, there may be quite a difference there. We are not just thinking about BacT/Alert but we are looking at any system and certainly the Pall system is an indirect measurement, looking at oxygen consumption, and we don't know that there isn't a difference. So, I would think that we would have to validate leukoreduced versus non-leukoreduced, although if you are looking at buffy coat platelets that are leukoreduced, I think those data are very applicable to our leukoreduced platelets. If you have data on buffy coat platelets that are not leukoreduced, those are very relevant to non-leukoreduced platelets.
DR. NELSON: Actually, theoretically non-leukoreduced blood might actually clear some bacteria that were present.
DR. KLEIN: I have heard that stated, and certainly the Europeans feel that their hold on blood for 18-24 hours prior to filtration is really beneficial but, again, I think we need some data on that and we really don't have any. We have numbers but we don't have data.
DR. DICKSTEIN: Rob Dickstein, Pall Corporation Yarrow. For the record, the Pall BDS bacterial detection system has been licensed in Europe for approximately a year and a half now.
DR. NELSON: So, yes, yes and no I guess.
DR. STRONG: Well, I think there is validation and then there is validation. There is no process that we change, as Jim points out, that we don't validate. I think the question here is do we have to revalidate the bacterial detection system itself rather than the process. So, we will definitely validate the process. Whether we have to go through the entire gamut of testing every organism at different doses, and what-have-you, I think is really the issue and then the differences I think are quite apparent.
DR. FITZPATRICK: The ultimate goal appears to be is, is there an impact on sensitivity of the detection system. So, I think each process or change in process needs to be evaluated on the number of bacteria that would be present at the end of the chance of the process. So, a non-leukoreduced product should have more bacteria present than a leukoreduced product. If there is an impact, then the manufacturer has to revalidate that system, but if it is validated on a leukoreduced product I wouldn't think you would have to go backwards.
I just want to support Dr. Klein. I think there is ample evidence there is not enough difference in the procedure in producing the platelet between Europe and hard spin to invalidate the bacterial detection systems or even scientifically think of a way that it could possibly invalidate those systems. So, that data should be very applicable.
DR. EPSTEIN: Just to play devil's advocate, if the white cells incubated overnight eat up more bacterial which are then cleared when you filter, then it matters how long you dwell and when you filter. In theory, you might have lower bacterial concentrations then that, depending when you culture, might or might not detect the contamination. So, the two things can interact.
DR. FITZPATRICK: If you are culturing at the end of that incubation period, theoretically you can do that, yes.
DR. EPSTEIN: Yes. What I am trying to say is you have to look at it as an integrated process. One of FDA's concerns about claims that bacterial detection systems are release tests, namely, that they assure at some level of confidence a negative culture at the time of issue or control testing outdate, has to do with when do you test and under what conditions do you test in relation to that assertion. It is the definition of those conditions in relation to the process as a whole that is the question.
DR. FITZPATRICK: I agree, Jay. That is why I think the analysis should be comparing the processes so that they are applicable, and if there is evidence to show that there might be a reduced load at the end of that period because of the white cell interaction, then the recommendation, if you are not going to revalidate, would be that you culture prior to that I guess to avoid that complication until you can figure it out.
DR. KLEIN: But remember that there is a lot of European data, and some of them are doing their pooling and filtering after eight hours, just as we do. Some are doing it, as the Dutch do, usually between 18 and 24 hours. Some are doing it at 24 hours. There are a lot of data out there and I would still argue that for any given process that is similar, whether you do a hard spin first or do a soft spin first, is not going to make a difference. You just want to make sure that the systems are comparable and then you can use the data that are generated in Europe.
DR. SCHMIDT: Does hard spin versus soft spin have anything to do with throwing down the bacteria into the final product? Are you spinning down bacteria?
DR. VOSTAL: Well, I don't have any data on that. I would think that platelets would be comparable size, maybe even bigger, so I would think that if platelets are staying up the bacteria are as well.
DR. LAAL: You would need to spin much harder to get the bacteria to come down.
DR. STRONG: Just to comment on the release issue, I don't think most of us are considering this to be a release test because you have to incubate it to get a final result, and if you incubate five days, you won't have any platelets left. So, the challenge here is at what point do you culture? When can you accept a certain level of detection? And, is that good enough?
DR. NELSON: After the culture is taken, how long does it have to incubate and be negative before the platelets are released? Twenty-four hours? I mean, sometimes you use platelets that are two or three days old.
DR. STRONG: I think the majority of people at this point are planning to store the platelets for 24 hours, take a culture, release the platelets, then allow the culture to go on as long as the platelet is still in storage. The data suggest that you should expect the majority of cultures to become positive at 24 hours of the initial culture.
DR. NELSON: At least with clinically relevant organisms.
DR. KLEIN: The Dutch are releasing right away, as soon as they take their sample, and then calling back any positives. Others are holding for anywhere from 12-24 hours where are virtually all the organisms that are going to show up will show up. But that is something where, again, I think there is going to be some controversy about when to do it, and there may be several ways.
DR. SIEGEL: I just want to mention that the two approved systems for quality control culture are different in that the BacT/Alert has a continuous readout, the culture continues to incubate and can be read out at any time during the storage period of the platelet. Whereas, for the BDS system you have a fixed time at which you read it, either 24 or 30 hours, but you read it one time and you do not continue to incubate. So, it is not exactly as Dr. Strong described if you use BDS.
DR. NELSON: Do you think most blood collection facilities are going to go to culture?
DR. STRONG: Well, we don't have a scientific survey at this point in time. I think it looks like the majority are but it kind of depends on your size as to which system works better for you, and also the logistics of how far you are transporting platelets to a test center.
DR. KUEHNERT: Matt Kuehnert, CDC. I was going to make a comment on the question about when to release platelets. I wasn't sure whether the question was when should people or when are people going to. It is sort of a difficult thing because of the spectrum of how organisms behave. Certainly, there is a bimodal distribution roughly between gram negatives and gram positives, with the gram negatives growing more quickly.
But one thing that was brought up earlier was endotoxin assay. That would certainly be a quicker potential way to capture units that might be right at the edge of becoming positive, had just been releases, say, on day two and then two hours later it gets recalled and it is sort of too late, already sent out. I guess that is not on the agenda but I just wanted to sort of add that as a potential ancillary test that could be used in conjunction with culture. But I don't think there is an easy answer to when to release because we are talking about organisms that all have different characteristics.
DR. NELSON: Are all potential blood donors who are on any sort of antibiotic, prophylactic or otherwise, excluded as donors, and is that system pretty good? Because this could influence how quickly an organism might grow out.
DR. KLEIN: Yes, they are unless they are on acne antibiotic but essentially they are eliminated because of their underlying illness and antibiotic is a marker for an underlying illness.
DR. STRONG: In terms of endotoxin, actually we are doing endotoxin assays now on certain cellular products, such as islets, prior to release. I think that would be a logistical challenge for blood centers who are processing 10,000 or 20,000 units a day to do that many endotoxin assays, but we have been forced to do other things.
DR. VOSTAL: All right, thank you very much.
DR. NELSON: Thank you. The next issue is an update on particulates in blood bags. Dr. Lewis?
Update on Particulates in Blood Bags
DR. LEWIS: Thank you, Dr. Nelson.
We appreciate the opportunity. This was a last minute addition to the agenda. Thank you, Dr. Smallwood for allowing us to put this on late in the planning process but it has been an ongoing concern at the FDA as well as at blood organizations, exactly what are white particulates that have been identified in some blood units. In order to update you on a number of ongoing studies, as well as the entire issue, we asked for some time here.
The FDA received information on this particulate matter from the American Red Cross on January 31st. Their report told us that they saw it in Baxter blood bags. They saw it only in the southern region and then, a few days later, noted that there were adverse events associated with some of the lots of blood bags in which they had identified particulates and that there was some investigation of a fatality of an individual who had received a couple of units of blood, and that blood was also collected in blood bags that were of the same lot as some of the others in which they had observed particulates.
This is one of the first photos that the FDA received. This is from Dr. Chris Hillyer at Emory. He had taken a unit of blood that had these particulates in it and passed them through a normal kitchen sieve and there is some material that didn't pass through the sieve. This was of some concern.
He has isolated one of these particulates and shows it here. It is about 8 mL.
During the course of the investigation, the American Association of Blood Banks was kind enough to put some photos of various types of particulates that had been observed on their web site. Throughout the course a number of different types of particulates were identified. Type C or type 1 is very small, later referred to as "starry night." Type 2 are larger particulates. Type 3, atypical bubbles were seen. Type 4, kind of a yellow oil slick on the top of the blood unit.
These are further pictures. In the upper right, a normal bag compared to a type 4, and then other normal appearing units.
This was type 5. Some of the particulates that are seen in the closeup on the right were later identified as materials on the outside of the bag.
It is funny now!
This is a close up of type 1. Some of the particulates there.
Type 2, with larger aggregates.
These were the oil bubbles. Initially the observations suggested that there was a particulate at the bottom of the bubbles, however, it was later noted that the nature of the bubbles was such that the light that was used to illuminate the blood bag was reflective in the back.
The same thing when the same bubbles were taken with polarized light. It was shown that there were in fact bubbles there that did not coalesce, however, there weren't particulates in them.
Again, type 4. This is the oil slick. I am not sure what the perspective is of this photo but you can see that in the middle there is a long, stringy fibrinous material.
From the Red Cross we heard that some of the conditions where they observed these particulates were that the bags had previously been in the cold, removed from the cold, the bags placed flat and particulates could be observed after about ten minutes. The observations were with the label down.
We noted that our regulations require a visual inspection and on February 7, one week later, the FDA issued a statement that we put on our web site encouraging visual inspection over and above the required visual inspection. We thought that the procedures that were used by the Red Cross were appropriate and we noted that this was an interim, precautionary measure.
At that time we asked that any observations be reported by email or by phone. One of the next speakers, Jerry Davis, from the FDA, will comment on some of the reports that we have received at those sites.
During all this time the CDC was involved and they did a number of studies. Matt Kuehnert is going to present some of those studies, both clinical as well as laboratory investigations into some of these particulates.
I am going to stop there right now. I have a number of other slides and I would like to finish at the end of our session with those particular slides, but we have a number of speakers who will tell us about some of the investigations that they have completed and that are ongoing. Dr. Peter Page, from the American Red Cross, will talk about the discovery of these particulates, some of the ADR investigations and the various conditions of collections that led to their conclusions.
DR. NELSON: This reminds me of an experience that I had as a hospital epidemiologist at the University of Illinois. There was a surgical patient that needed a unit of fresh-frozen plasma immediately. So, the junior resident went to the blood bank, got the fresh-frozen plasma and on the way to the OR stopped at a food microwave, turned it on, and the bag partially exploded. I was called because of the blood on the inside of the microwave. But it was transfused into the patient and it was highly particulate at that point.
Discovery, ADR Investigation, Conditions of
DR. PAGE: This will be a little different from that.
I have a number of slides presenting our initial observations, actions and thought processes as we went along, and I am going to try not to duplicate what I expect will be covered by others.
The first couple of slides are background information which will help put in perspective some of our earlier findings and rationale. The grey parts of the map represent the 36 Red Cross regions in which Red Cross traditionally recruits, collects and distributes blood components. The 36 regions are divided into four areas, as indicated on this slide, for management purposes.
This slide is the same map but reflects an initiative that was ongoing in January and February of this year. The American Red Cross uses all three vendors' blood collection sets for whole blood, the primary one of which is Baxter. We were in the process of converting from an older Baxter blood collection set, known as SampLink, to a newer Baxter whole blood collection set, known as Y-Diversion. The three areas, western, north central and north Atlantic, had completed their conversion to the new collection set. All the leftover bags, in a variety of lot numbers, from the older collection set had then been sent to the south central area to be used up. So, all the older Baxter blood collection sets were in the southeast of the U.S. and represented a larger number of different lots than ordinarily individual regions would use.
The first region in which white particulate material was noticed is the southern region, which is based in Atlanta, and the next region in which it was noticed was the Tennessee Valley region, which is headquartered in Nashville.
Red Cross collects over six million units of blood using Baxter whole blood collection sets for two-thirds, Pall/MedSep for a quarter and Terumo for eight percent. The southern region collects essentially only in Baxter blood collection sets and was using all of the older version. The southern region is not able to provide for all its hospital needs from its local collection but imports 44,000 units of red cells from other regions which could use different lots and a variety of blood collection set manufacturers.
The Tennessee Valley region is a smaller region and uses Baxter, the old collection set, but 38 percent of their blood is collected with Pall/MedSep. They collect more than their hospitals need so they export units. These differences had an effect on how the impact of what we did affected those two regions.
This is one of two slides giving day by day chronology of the early events, a couple of which Richard Lewis already referred to. It was on Monday, January 27 when a new employee noticed something that he thought might be peculiar in a unit of red cells and brought it to his supervisor's attention. A number of other experienced blood center staff looked at it as well and agreed that this white glob or particulate material was unusual. So, more red cells were inspected and more units with white particulate material were seen, and all those units in which it was seen and the plasma corresponding to it were quarantined. That day they began an inspection of the entire red cell inventory in the Atlanta region for this.
On Tuesday, the next day, more units were seen with white particular material in the inventory and those and their co-components were quarantined. If what particulate material was seen in a unit of whole blood or if it was seen in a red cell prior to leukoreduction processing stopped. In other words, we did not continue our ordinary practice of separating whole blood into red cells and plasma, and if we saw it in a non-leukoreduced red cell we didn't leukoreduce it. We weren't going to be using the unit or any of the components in any event. That was the day that we notified the bag manufacturer, which in this case was Baxter.
By the next day we had noticed 50 units of red cells with white particular material from a multiple number of different bag lots, all the old Baxter collection sets. Chris Hillyer, at Emory University, in particular began looking at these after straining in the kitchen sieve and also under the microscope in a variety of ways, which we will talk about later.
Next day we saw 80 involving even more bag lots. So, Red Cross quarantined all the southern region collected units in lots that had been affected to date and provided only red cells imported from other regions to its hospitals. The Red Cross asked hospitals in Georgia to not transfuse souther region locally collected units except in emergency, if they were needed, and in that event to inspect the units prior to transfusing the red cells. If the unit was collected in a region other than southern region, it was okay to use those units. Accordingly, a large number of elective surgeries were cancelled in a large number of hospitals in Atlanta that day.
Next Friday the number continued to grow as we continued to collect and were noticing it in more bag lots as all collections were being inspected. That day we alerted the FDA and that same day, in Atlanta, the FPPI came to pick up a couple of units of red cells with white particulate material to deliver to the CDC for testing, which you will hear about later. Due to the lack of surgeries, being postponed, and the blood shortage there was a press conference.
On Saturday Chris Hillyer called me to say that there was a young patient on ECMO, extracorporeal membrane oxygenation. It replaces the lung for patients with severe lung disease. An experienced ECMO tech, whose seat is right in front of the filters in the machine, noticed white particulate material accumulating in the filters. The process went on for several days so we weren't able to remove the filter to look at it yet but this tech said these filters were in front of his face every time he did an ECMO, which he had done for years, and this was new; he had never seen it before.
On Sunday, at 1:45 in the morning, we got a call that the Tennessee Valley region had seen white particulate material in ten units in using four bag lots. Shortly thereafter, it was 23 lots and some 30-some units. At nine o'clock that Sunday we spoke with the FDA and there was a lot of discussion as to whether and what to do since this was felt to be a new finding of unknown significance and we didn't know for sure what it was that was in these bags. We were somewhat reassured by the fact that there had not been any adverse reaction reports associated with this to date. But we discontinued use of the older bags in all regions, all the regions in the southeastern U.S. We precipitated our conversion to new bags and got new bags into all those regions and discontinued use of the older bags. There was one day on which we didn't collect blood in two regions because we didn't have staff trained for the new bags in tie but we got that solved by the next day.
We quarantined all the components, red cells and plasma, from all units collected in all lots of older bags in the souther region and Tennessee Valley region. There was a lot of discussion about this. We didn't think it was lot specific because of the large number of lots involved, and we weren't prepared to, or FDA agreed there wasn't a rationale to require doing this for the whole system, but we would focus the initiative to the two geographic sites at which this had been observed since it had not been observed elsewhere yet, thinking that there might have been something different in those geographic sites.
We then began an inspection of our entire inventory in all regions and we initiated inspection of all upcoming whole blood collections. The unit that the CDC had received--there was a report on biothreat agents that they have a standard screen for using PCR and it was negative, permitting the CDC to then have their environmental and chemical lab begin its analysis.
Next day is Monday. Baxter and Red Cross scientists from the Holland Lab and from Chicago were in Atlanta that morning, examining the white particulate material in some detail.
On Wednesday--well, let me say that Red Cross had earlier reviewed the adverse reaction reports from Tennessee Valley and southern regions and had not found any reports since January 1st that seemed to be temporally or causally related to the timing of our observing these white particulate materials. However, after we had asked the hospitals not to transfuse any southern region collected units and to quarantine them, a physician in the hospital gave second thought to a patient who had died about ten days earlier who, in retrospect, died while receiving a second unit of red cells. At the time that unit was not felt to be a transfusion reaction related to the death but, based on thinking about it further, they gave it second thought and raised the question. We were notified. We notified the FDA that day. We made the decision later that day to withdraw all components collected in those two particular lot numbers of bags system-wide.
As Dr. Lewis mentioned, there were two bag sets involved in those units of red cells transfused to that patient, and one of the lot numbers had already had 18 other units with white particulate material identified in its collections. The other lot number for the other red cell had one other one. So, we thought it would be prudent to take that initiative at that time.
On Friday, February 7th, we noticed white particulate material in another vendor's collection set and we also noticed it in units collected from Baxter's new whole blood collection set. We got reports that non-Red Cross blood centers had noticed white particulate material, some of them, involving the third collection set vendor as well.
This is sort of a summary of the initiatives we took going back to January 27th, when we right away quarantined the red cells and plasma from each whole blood collection in which white particulate material was seen in the red cell. Then, on February 3rd, we quarantined and withdrew all components collected in the southern region in all Baxter collection sets. A couple of days later we did the same thing in the Tennessee Valley region and Nashville as well.
After the fatality, which was at least temporally related to transfusion but not necessarily causally related, we withdrew all components collected in the two collection set lot numbers used for transfusion to that patient from all of our regions. Those were primarily in the southeast.
Then subsequently, a bit later, on the 10th after some discussions with the FDA and based upon some safety studies which were not alarming, which I will get to in a minute and you will hear more about later, we released from quarantine all the leukoreduced red cells, quarantined up here. Remember, that was almost 10,000 red cells at a time when the blood supply was not flush. I may be getting to that in a minute.
The plasma units we kept in quarantine and we destroyed any red cells that had not been leukoreduced. You will hear later about the number of investigations that we have started and are ongoing.
The next several slides are a variety of pieces of information. In Atlanta we quarantined 7,869 red cells. I don't have on here Tennessee Valley. It was a substantial number but not as large. There were more than 10,000 plasmas involved, plasma having a longer shelf life, units collected in those lot numbers more than 42 days ago were still in inventory and that is why the number is bigger. These numbers represent, on a day by day basis, the number of red cells that we shipped in from other blood centers to supplement the blood supply in Atlanta since the quarantine had had a massive effect on them. So, there were 7,800 quarantined. We shipped in almost 9,000 units. We want to express thanks to several other non-Red Cross centers who helped out in sending several hundred units as well. They were more than gracious and offered to help. These are the number of plasma units that were sent in, including plasma from a non-Red Cross center as well.
This is that same map again. The red slashes indicate regions in which Baxter blood collection sets, the old and the new, had had white particulate material found. Regions of Pall only are blue, and the combination of Baxter and Pall. So, there was no geographic focus or logic that we could make out of this although we did look.
These are product code numbers, and these are anticoagulants, and these are the component packs, whole blood, red cells, red cells leukoreduced from Baxter and Pall, and the number of units with white particulate material seen in them. The total here is 356. I would like to point out that it was seen in 16 units of whole blood, that is, units that had not been spun at all and units that had not been filtered. It was seen in 286 non-leukoreduced red cells, but it was also seen in 70 leukoreduced red cells. Remember that if we saw it in whole blood we stopped processing and didn't separate it. If we saw it in a non-leukoreduced red cells we stopped processing and didn't leukoreduce them. The point here is that it was seen in leukoreduced and also prior to leukoreduction or prior to separation.
Dr. Lewis gave you nice pictures of the four types. Red Cross has not seen any type 4s and we have seen basically half and half type 1s and type 2s in our first several hundred units.
This tries to describe the frequency of the observation. This goes back a while now. In the southern region, in the first 11,800 units we inspected we saw 198 units with it, for a 1.6 percent observation rate. Tennessee Valley inspected about as many but saw fewer, for an 0.28 percent rate. Nineteen other regions had 0.01 to 2.15 percent observation rate so Atlanta was not the highest. In 15 regions we saw none in over 40,000. So, in about a quarter of a million units we had seen 662 and 0.28. An updated figure from Atlanta alone is that as of ten days ago we had seen 973 units, representing 53 lots, 49 Baxter and 4 Pall MedSep.
Putting together the recipient complication reviews that we had, Dr. Linda Chambers, of our donor and recipient complications program, reviewed from January 1st from all 36 regions and there were 45 cases excluding infectious complications and of all levels of probability of association with transfusion, in other words highly suspect and not so highly suspect, and including regional investigations that are not yet complete. Sixty-five percent of the involved components were collected in Baxter sets. It is actually 66 percent of whole blood collections for Red Cross use of Baxter sets so there was no over-representation of white particulate material in Baxter sets.
Nineteen cases were submitted as possible TRALI, transfusion related acute lung injury. Historically, we have good data on this and get 2.4 cases per week. So, in 7.5 weeks we would have expected 18.3 cases and we had 19. Nine were in patients who received blood only collected in Baxter. Another nine involved only non-Baxter collections and one involved both. So, there was no over-representation of the observation in Baxter collections. This took us a little while to compile.
I put in this slide because it is an interesting observation that I don't yet understand. There was a high school blood drive in Georgia using a single Baxter old blood collection lot number. These are the whole blood numbers, unique number assigned to each collection. They are all serial in a row. There is one missing here but that was QNS. That was a unit that was incompletely collected so there wasn't enough blood in it for making a component.
We saw on several blood mobiles runs of serial whole blood numbers. We did also see one by itself or two or three, but it was striking that some others would have none or one mobile would have a long run, a gap, another run, another gap. There were type 3, type 2 and type 1 in this run. This was especially where we were monitoring a large number of other parameters regarding that blood collection mobile unit--you know, how many minutes did it take for the unit to collect? What were the lot numbers of the arm preps? Was there a particular phlebotomist involved? Was there a particularly newly trained set of phlebotomists involved? The answer is we found no variable that we could tie to it.
There was a large number of studies, and I am sure this is not complete because I know Baxter, FDA, Emory and others have also done this, but I have divided them by the studies on the anticoagulant preservative in empty collection sets, and then examination of white particulate material itself, and then red cell units that have material in it. So, empty whole blood collection sets from implicated and control lots were frozen and thawed and no particular material was seen in these solutions. Units were then collected at the Holland Lab in these sets and no white particulate matter was seen.
You might wonder why we did this. There was a particularly cold snap of weather in Atlanta right prior to the initiation of this observation, as well as there was in Nashville, and this was an early postulate and concern. The anticoagulant preservative solutions were filtered and no white particulate material was seen on filters by microscope, the question being were they in the bag before blood was collected. The CDC's environmental and chemical lab found no difference in the number of volatile organic compounds in empty blood collection sets from implicated lots versus control. Emory, Baxter, Red Cross and others have done staining of peripheral smears from units with white particulate material. There have been traditional paraffin block sections. There have been a number of studies which I suspect will be described later. None of these showed any foreign or unexpected material in them. The material is largely platelets.
If you look at red cells with the material, the biothreat assay was negative, as I mentioned, by CDC. Their culturing is negative for bacteria and I think microbacteria may still be pending but everything has been negative. The CDC assay for endotoxin was negative. They looked at red cell units with and without white particulate material for volatile organic compounds, but we later figured out that the control units were collected in non-Baxter bags so that the results were no invalid; there were other reasons for the bag to show the differences, but nothing of concern was identified there.
FDA and we are working on looking at pre-storage leukoreduction and standard blood administration set filtering to see what particulate material is reduced or eliminated by pre-leukoreduction filtration, and then the blood administration set that is used at the bedside as all units of red cells are transfused.
I won't go through these but I alluded to it on an earlier table. The manufacturing variables with no correlation are all these things which have been looked at informally, some; others more formally and those informal to date are going to be looked at rigorously by CDC statisticians, whose help we appreciate.
Those variables with correlation, platelet rich components. In Atlanta we don't make platelets from whole blood collections, and it was seen primarily in units from which platelets were not to be made, which meant that the whole blood was shipped to the lab in the cold and the only spin was a hard spin, packing the platelets down in there. It was seen pre-leukoreduction. If you then leukoreduced it larger materials didn't go through and there weren't as many of them seen. Initially it was correlated with Baxter collection sets but remember that Atlanta only used Baxter collection sets so that did not pan out.
I mention this in part because we are reviewing temperature and time of transport from the collection site to the component lab. This is not final, but nothing striking has appeared to us yet. The co-component plasma, in view of the fact that these were felt to be adherent groups of platelets, some but not all of which appeared activated, it has been suggested and we are arranging to have performed the plasma in those units to be evaluated for products of platelet activation and also products of the clotting cascade. There has been no evidence of clots in the red cells or no evidence of clots in the plasma upon traditional visual examination.
The fatality is felt to probably now be unrelated. The review of recipient complications from all regions has been negative and not shown an increase recently or association with any bag type. I believe Dr. Kuehnert will tell you about the Georgia Department of Health survey of all hospitals which did not find anything striking or surprising. Two large Boston hospitals which both collect blood themselves and receive blood from the Red Cross had noticed white particulate material in both types of components. They reviewed their transfusion reactions for the last three months and compared it with the previous year and did not find any uptake or increase.
This is my next to last one, I believe. The overall assessment--I think it is important to point out that there is no evidence of foreign material in the components but we did not know that in the beginning and we did not know it with assurance for a while. As I think you will hear, they are primarily platelets, some but not all activated. The size and numbers seem to be decreased by leukoreduction filtration. No correlation with adverse reactions, thankfully.
It is not clear still why it is newly noticed now. A number of long-time people who work in labs say if it had been in there before, particularly the type 2s, they would have noticed before and a number of people feel it is new. The investigation is ongoing. There are a number of studies still in progress.
My last slide just sort of summarizes Red Cross recent actions. FDA did come out with a statement near the end of February, and it was around that time that we released from quarantine. But prior to release from quarantine we re-inspected every unit and we did not release any unit in which white particulate material had previously been seen. We kept in quarantine or destroyed all components from any donation with white particulate material. We left standing the voluntary withdrawal of all the components from the two collection sets related to the fatality. We did not release any non-leukoreduced red cells, of which there are very few because we use essentially universal leukoreduction. The plasma, as I mentioned, remains in quarantine pending the coag. studies. If they are negative, we have an internal process of material review board which would permit us to review and consider whether they could be released.
We plan to reassess as new information accrues and when all studies are completed. We appreciate the help and cooperation of a large number of other agencies, the FDA, the CDC, Baxter and Emory, and look forward to continuing sharing of information and collaboration in this regard. That is it. Thank you very much for your attention.
DR. NELSON: An interesting investigation. How old are the old Baxter bags? Have they been around for quite a while?
DR. PAGE: The dating period on collection sets I think might be on the order of 18 months. The date of manufacture of these bags went from February, 02 up through November of 02. So, the oldest would have been almost a year.
DR. ALLEN: But well within the time.
DR. PAGE: Oh, we do not use outdated blood collection sets.
DR. STRONG: Did you notice any difference between collections on mobiles versus centers, or what is the mix there?
DR. PAGE: No, we didn't. We have a number of fixed site centers in Georgia. One uses automated mixers and that is the only one. They saw some but no more than others. We saw it in most of the other fixed sites. There was no correlation with the temperature of the weather, the day of collection, the distance from the center. We looked at everything for which we could get data.
DR. STRONG: I just noticed that in that sequential numbers that you had, that looked like a high school drive and depending upon how many staff and how many donors are showing up, they can get pretty hectic so sometimes they under-mix or don't store the units in the right way and it is very difficult to correlate any of that.
DR. PAGE: I agree. Our traditional staffing for high school drives is we send one extra nurse, just because it is a high school. We had collection staff and component staff from other regions come visit Atlanta and look over their shoulders, review their procedures to see if anything had been changed, seeing if they weren't following the procedures, and mixing was looked at carefully and did not seem to be different.
DR. STRONG: I might add though that when you have somebody looking over your shoulder you tend to do a much better job. We have seen those kind of variables on mobiles.
DR. PAGE: I agree; I agree. I would say that this was also observed in non-high school drives. There were a couple of high school drives in the first week but it was observed at fixed sites which probably are the most controlled environment, and it was observed at a number of drives that weren't high schools. I just picked one as an example.
DR. NELSON: The makeup of these bags I guess is proprietary information but I wonder what the differences are between the old and the new or between Baxter and Pall. It is probably a complex plastic with many different components.
DR. PAGE: The Baxter old and new, there is no difference in the bag. The only difference is the plastic on the tubing from which you collect the samples for testing. There are differences between Baxter and Pall, and those were provided to CDC as they were doing their analysis of organic volatile compounds. I think Baxter is speaking later and could talk to that better than I.
DR. GOLDING: I appreciate that this is an ongoing investigation and we are not sure which types of particles are more worrisome than others, but just looking at the pictures that Dr. Lewis showed, some of then are probably not significant. In other words, if you see bubbles, and you have different lighting and you stop seeing the bubbles that particular particle is not important.
My point is that once we have defined which of the particles are the more important particles or more likely to be pathogenic, in other words, if you see platelets or you see fibrin or you see something that looks more suspicious, it may be worthwhile considering going back and looking at the data and doing some kind of subset analysis in terms of if you now look at the ones that look like they could be pathogenic, what was the adverse event rate in terms of those particular bags, and other subset analyses that may be relevant in this situation.
DR. PAGE: I am not sure I understood your train of thought but I don't believe we have seen any adverse reactions that could be tied to this yet. I mean, there are those who worry more about the type 2s in as much as they are larger.
DR. GOLDING: Well, for example, you showed the incidence of TRALI and you said, well, the incidence wasn't different with those particulates or without the particulates. But if you now just looked at the incidence of TRALI with particulates that were composed of platelets, the type 1 or the type 2, and you now look back at acute lung injury, will something drop out of that that is very different? At the moment it is confounded by a lot of particles that probably have no relevance to the safety of the product.
DR. PAGE: I should point out that we are not aware of a unit with white particulate material observed and having been transfused. Now, that could have happened prior to our noticing the problem, but when we did we started inspecting all units and told hospitals to inspect any that they used. But there was no difference in units transfused collected in the old Baxter bags of similar lot numbers. So, I don't think we know of a unit with white particulate material having been transfused.
The other point I might just make is there are those who say they have been there all along. The literature and older textbooks do describe small white particulates being seen in blood after a week or a couple of weeks of storage. These were seen in whole blood, fresh from the blood mobile or the blood center, and were seen before and after leukoreduction, which is done in the first day or so after collection.
DR. ALLEN: I had dinner last year with a blood banking colleague who related to me that they had had a donor at their center who was on the apheresis machine after having eaten two Crispy Cream donuts. The level of chylomicrons was so high that it totally plugged up the machine. We may need to start taking food histories from our donors. Just in terms of full disclosure, I own no Crispy Cream stock. I do eat Crispy Cream donuts.
DR. NELSON: Which is why the high school is so suspect I guess.
DR. PAGE: In platelet or plasma apheresis you can notice lipemia during the collection. I whole blood you might not. But in the sample tubes that are sent to the laboratory for testing after spinning visual inspection is required and testing is not performed if it is lipemic. That is found, but with a low enough frequency that it hasn't been justified to ask donors that question.
DR. LEW: I am just curious since it looks like it might be platelets or you all are thinking it might be, have you all looked at the platelet counts in the particular particulate positive blood versus no particulate positive blood?
DR. PAGE: We haven't. This is all in whole blood collections and we don't do platelet counts on whole blood donations. When whole blood came from the mobile and it was seen there, one could arguably do a platelet count on those but we didn't, and we weren't thinking about it at the time. I don't believe we know well what the platelet count should be on a red cell unit pre- and post-filtration for leukoreduction, but that is something that is being looked at prospectively now.
DR. ORTON: I am Sharyn Orton, from the FDA. I am going to be giving a presentation in a few minutes that is actually going to address some of that because I do have some data from other blood centers.
DR. NELSON: Good. Thank you. Next is Jerome Davis.
Chronology and Field Overview
DR. DAVIS: I have some slides.
I am in the Office of Compliance in the Center for Biologics.
I am going to talk briefly about FDA's investigation of the reports on particulates. Dr. Page gave a very good chronology already and so I am going to try not to repeat too much of what he said, but I am also going to give an overview or kind of a summary of the particulate reports that we received.
As has been stated already, CBER was notified on January 31st by the American Red Cross that the southern region, Atlanta, Georgia had been finding multiple units of blood products that contained white particulates. The bags used to collect these units were made by Baxter and multiple lots of the PL-146 bags were involved.
We learned that as the Atlanta Red Cross suspended the use of certain lots of Baxter bags and put new lots in use, they began finding particulates in those additional lots. The number of bags with particulates and the number of implicated lots of Baxter bags continued to increase over time.
On February 2nd, 2003 ARC's national location, the Tennessee Valley region, also reported finding particulates in units from four lots of the same type of Baxter bags that were being used in the Atlanta region.
A number of hypotheses were posed as to how those particulates were forming. It was suggested, as somebody mentioned earlier, that the unusually cold conditions in Atlanta this winter may have caused inadvertent freezing of the bags. Our follow-up investigation at the American Red Cross facilities focused on the receipt, storage and handling of the blood pack units both in-house and during mobile collections. We also focused on the follow-up investigations that were done by the American Red Cross.
In addition, official samples were collected of units that the Red Cross had observed particulates in, as well as unfilled Baxter collection bags. Other Red Cross locations, including Birmingham and Mobile, Alabama, eventually also detected particulates in units and FDA initiated inspections of those sites also.
On February 3rd FDA began receiving reports of particulates in blood collected by non-ARC facilities. In addition, FDA began to receive sporadic reports about particulate matter in bags manufactured by Terumo and later on in bags manufactured by Pall. FDA also conducted inspections of the blood bag manufacturers, first Baxter and then eventually Terumo and Pall. Those inspections focused on complaints, adverse reaction reports, bag components and materials, manufacturing records and their follow-up investigations regarding the particulate reports.
As FDA's investigation continued, additional samples of blood units containing particles, as well as unfilled blood pack units, were collected from the American Red Cross and from non-ARC collection centers and investigations were conducted at some of those non-ARC collection centers.
We also sampled bag materials and components for analysis. At last count, 13 of FDA's 19 district offices were involved to some extent in the investigation, and this continues to be an open and ongoing investigation.
On February 7th FDA issued a "talk paper" and an information alert on particulate matter in blood bags. In both documents the agency recommended that blood establishments implement the ARC's enhanced visual inspection procedures and, if abnormalities were detected, to quarantine the products and report their findings and any adverse reactions potentially related to the presence of particulate matter to the FDA.
On February 27th FDA issued another press statement and I think that has already been mentioned. We also issued this frequently asked questions document which answered questions that were being posed by the blood industry regarding the management of units containing particles, as well as their co-components.
I wanted to go back and talk about the February 7th "talk paper." That requested reporting of abnormal units and any adverse reactions that were potentially related to the presence of white particulate matter. We requested that the reports be reported to the BPP Deviations email account and also to this 800 number.
CBER set up an access database to capture all the particulate reports and the adverse reaction reports or, you know, any that happened to be reported.
As of March 10th, we had only received one adverse reaction report, and this was a rather vague report where a woman indicated that her mother-in-law had had difficulty breathing post-transfusion. We had attempted to contact this person to get some additional information. I think this was actually reported on February 10th and up to this point in time we still have not received any additional information. There still really isn't any real link to white particulate matter. Like I said, it is kind of a general complaint. We didn't have any AERs reported to the 800 number.
As of March 10th, we had received a total of 97 reports. Actually, not all of them came into the email address or to the 800 number. If we received reports while our investigators were out inspecting the bag manufacturers, if they had received any complaints or any reports of white particulate matter, that information came back to CBER and we also contacted those establishments to collect information about them. There were also some other reports that came in to other parts of FDA.
Like I said, as of March 10th we had received a total of 97 reports and 94 of them were collection facilities and three of them were transfusion services.
We attempted to get kind of standard information from all the reporters. Some of the key information we got was the number of units with white particulate matter, as well as the number of units that they inspected. We tried to get an indication of what types of white particulate matter they were finding and in what products. We wanted to collect details about the blood pack units; who was the manufacturer; the lot numbers of the bags; how the bags were stored and information like that. We wanted to collect information about the processing of the units, were they refrigerated or not; were the units hard spun or were they soft spun; were platelets made, that sort of information. We also asked what was the method of agitation.
This slide gives just a general idea of the geographic breakout. The 97 reports were received from facilities in a total of 40 states and the District of Columbia. As you can see just from the states that are listed there, it is a pretty widespread distribution from coast to coast, and north, south and the center of the country.
In the 97 reports there was a grand total of 1,508 bags that contained white particulate matter. There was a grand total of 315,000-plus bags inspected. That gave us a percentage of bags with white particulate matter of about half a percent. I notice that differed somewhat from what Dr. Page reported. I think he said the Red Cross had 0.28 percent.
This gives a general breakdown of the bag manufacturers. The majority of the reports were in Baxter bags but, as you can see, there was also particulate matter found in the bags made by other manufacturers. In about 20 percent of the cases we didn't get the information reported.
This slides is a breakdown of the types of products that we found particulates in. The majority of them were red cells. The second line there, you can see these are leukoreduced. In 103 cases the particulates were found in red cells that had already been leukoreduced. These are autologous red cells and in some cases autologous whole blood. There was only one case where there was a leukoreduced whole blood unit. Quite a few of the reporters did not report this information.
This gives a breakdown of the types of particles that were found. The great majority of them were types 1 and 2. Again, this refers to the types that are on the AABB web site, the types that Dr. Lewis went over. I think Dr. Page said they didn't see any type 4. We didn't see much. It was interesting that in some cases reporters found particles but they didn't feel like they fit into any of these four categories. Again, in about 20 percent of the cases the reporters did not indicate what type of particulate matter they were.
This gives and indication of how the units were processed. Again, out of the 97 reports, in 26 cases the unit was refrigerated and then hard spun. Down here, in 32 cases the unit was hard spun but there was no indication given by the reporter whether or not the unit was refrigerated or not. But you can see that in the majority of these reports the unit was hard spun. Here we only had five cases where the reporter did a soft spin and made platelets. Again, in about 20 percent of the cases the reporter did not indicate how the units were manufactured.
This is the last slide I have. There were 94 collection centers. Kind of late in the process of gathering this information we started asking about the method of agitation and in quite a few cases we were not able to get this information, but when we did, in 28 cases the method was manual. In 20 cases it was automated. In four cases it was a combination of both.
That is all I have unless there are any questions.
DR. NELSON: Thank you. Were any of these reports made during or after transfusion into a patient or were these all units that were noticed prior to transfusion?
DR. DAVIS: I believe that they were all reported prior to transfusion.
DR. NELSON: From what we saw, presumably there must have been a fair number that were not noticed and were actually transfused if the prevalence is half a percent, I would suspect. Yes?
DR. PAGE: With respect to the prevalence, the difference between the 0.4 percent as reported by the FDA and the 0.28 percent reported by me and Red Cross, I think there are two factors there. I believe, from listening to you, your prevalence is based only upon blood centers that saw any. Of Red Cross 36 regions, there were 15 regions that did not see any but their inspections were included in our denominator which would accordingly be a basis for being larger.
DR. DAVIS: That is a good point.
DR. PAGE: Also, in the beginning we did not report these as real until they were "confirmed" by observers in Atlanta experienced in these observations. Particularly prior to the AABB web site showing pictures, we wanted to make sure there was consistency in what we were reporting, measuring and acting to. So, there were a fair number of suspect units observed in other regions, sent to Atlanta, that in Atlanta were declared to be not real or not it. So, I think we may have prevented some over-reporting of false positives.
DR. FITZPATRICK: What is the current level of reporting in the past? Is there a bell-shaped curve in reporting? Are we seeing less reporting now than we had?
DR. DAVIS: Yes, the number of reports has definitely fallen off over time. I think since last Thursday, eight days ago, we have received one report.
DR. FITZPATRICK: And the distribution of reports of bag manufacturers, is that about proportionate to market share?
DR. DAVIS: I can't answer that. Maybe Steve Binion can address that.
DR. NELSON: Who is the next speaker? Steve Binion, from Baxter?
DR. PAGE: From the Red Cross experience, remembering that we use all three vendors' bags, Baxter is right on the money as to percent of white particulate material observations compared to percent of bag collections. The other two are pretty similar to the percent of bags. So, I wouldn't say that there is a real difference in frequency.
DR. BINION: Steve Binion, from Baxter. I am just glad that Dr. Page was able to handle that because I don't have the market breakdown with me.
DR. NELSON: You should! Steve, are you going to make a presentation? You are listed.
DR. BINION: Yes.
DR. NELSON: We are ready for you.
DR. BINION: Steve Binion, from Baxter. I apologize, I was looking at the wrong agenda I guess; I thought there was another FDA speaker.
DR. NELSON: Oh, right. The nameless person from FDA, Betsy Poindexter. Sorry, you weren't listed. Go ahead.
Testing, FDA Findings
MS. POINDEXTER: Good afternoon.
I am here today to present not so much the FDA findings but those studies that are under way by groups that have agreed, either coercively or voluntarily, to perform studies for us.
I am with the Division of Hematology, Laboratory of Cellular Hematology. The labs that have volunteered to do some studies for us, in addition to our own lab--I am with Dr. Vostal in the Division of Hematology--are Dr. Joe Hutter, at CDRH in the Office of Science and Technology. They are looking at both the bag surfaces and the bag composition, as well as the anticoagulant solution. They don't have the facilities for handling biologics so they won't be handling any of the blood products per se.
The Forensic Chemistry Center in Cincinnati, Dr. Karen Wolnik and her group are looking both at the blood bags, the anticoagulant solutions and particles that have been in those bags. They do have blood products.
Dr. David Stroncek, at the NIH, Department of Transfusion Medicine, has voluntarily submitted a protocol to us and we have commented on it. He believes that in the next couple of weeks he will be able to initiate his studies with one of his SBB students to try to replicate this phenomenon in their hands, using bags that were collected from the field as well as the bags that they generally use in DTM.
Dr. McDonald Horne, at the NIH Clinical Center Hematology Service, has worked with Dr. Stroncek and is working with us to identify some of the fibrinous like or particulate like material.
Dr. Gary Moroff, at the Red Cross, has also submitted a protocol for trying to replicate these procedures at the Red Cross at the Holland Labs.
In our laboratory at CBER we are looking at visually examining the blood products. We have taken digital photographs of each of the products that have come in at various time points so, if there were no particles to start with and particles developed over a period of time that we have had them in refrigerated storage, we would capture those events. We are also distributing the blood products that have been gathered from the field to the test facilities that have been doing some testing for us. We are collating the test results. Right now that spreadsheet is blank because the tests are in very preliminary stages, and we do have a database for all the samples that we have collected, all the blood unit numbers, the lot numbers of the products, the bag manufacturers, physical descriptions of bags, the collection sites, etc.
These are some photos that we have taken of the blood units that we have received. As you can see on the left side, these were definitely type 1 and perhaps type 2 particles. These were the "starry night" or dandruff-like particles that were described in Dr. Hillyer's photos. They were not there when you first got the bag out of the refrigerator but if it sat there for a few minutes they rose to the surface. If you were to mix the bag these would go back into suspension but then if you were to let it settle again, they would again rise to the surface. The bag on the right-hand side, again, has flake-like, type 1 material.
If you notice this blood, here, if you look at this unit, let it sit long enough you see the fatty plasma settle out and you will have this white sheen across the top of the unit. So, there is the blood here and then there is a little bit of a surface there. The bag on the right-hand side was a unit that was collected through a military facility. You can see that it has the bubbles and it has the reflective pattern in there, as well as some little particulate matter in various areas of the bag.
It is a challenge to try to photograph these things because if you use a flash on your camera you create so much more reflection than what is already being created by the overhead lights, as you can see from the adhesive tape on this bag, but you can definitely see particulates in various areas of this bag, and platelets probably would not have been made from this unit. I should probably put type 1 and type 2 for this bag, as well as reflective bubbles.
What I might point out on this one, as well as this one, is that the material was adherent to both the top surface and the bottom surface of the bag. So, irrespective of whether you turned it label side up or label side down, the material was adherent to the bag and even with gentle agitation it was unlikely that that material was going to move around.
At the CDRH, they are going to employ a number of different tests for the bag materials, as well as the anticoagulant. They are using differential scanning and calorimetry, gel permeation, chromatography, FTIR to evaluate anomalous peaks in the PVC used in the storage container. HPLC may be used to evaluate adhesives on labels. Soxhlet extraction matters for plasticizer analysis and thermal gravimetric analysis to analyze material stability after extreme temperature storage.
We have provided CDRH not only with some of the test bags and the control bags, meaning a lot that was manufactured this year rather than last year at the Baxter facility, but we took bags that we had in our lab and subjected them, similarly to what the Red Cross did, we froze some and thawed them. We also heated some to 56 degrees for a number of hours and then brought them back to room temperature and froze them. Early in this investigation we were very concerned that there may have been some transient thermal effects or cold effects on the anticoagulant or bag systems that might have caused these problems to occur in the Atlanta region. So, we were sort of investigating all possible realms.
The FCC or the Forensic Chemistry Center in Cincinnati, is looking at the anticoagulant solutions, the blood storage container. The are also using FTIR and GS mass spec, SEM and microRaman spectroscopy, as well as light microscopy. They do have blood products with particulate matter and they are doing photography, chemistry and histopath. analysis on those samples.
This is a photo that they sent us through email last week. This is a particle that they pulled out of one of the blood bags that they had. As the red arrows generously point out, you can see the cross-hatching on this particle that is identical to the cross-hatching on the internal surface of the red cell bag. When I saw this photo I said great shot but, number two, I wanted to know whether this was a fee-floating particle or whether this was a particle that was adherent to the inner surface of the bag. The scientist said that it was something that he was able to pull out of the blood; it was free-floating in the blood. So, it was obvious that at some point it had laid on the surface of the bag and was probably adherent. Maybe through additional jostling of the bag, through the manipulations it was feed up but still maintained that scar of contact.
At the NIH Department of Transfusion Medicine they have developed a protocol to compare the test bags recently evaluated to the control bags of another manufacturer. They are evaluating donor characteristics such as diet, over-the-counter medications or prescription drugs, blood chemistries and complete blood counts and coagulation profiles to identify perhaps if a donor whose blood goes on to form particulates may have a higher fibrinogen, a higher platelet count, a higher white blood cell count to start with that may just aggravate what may be a rather normal occurring phenomenon.
They are also going to look at the pre-storage leukocyte reduction on the formation of these products and see whether, if you filter fresh after collection within the first couple of hours at room temperature, whether you can perhaps reduce the numbers of particulates that are formed versus leukocyte reduction within the 72-hour period that is generally allowed for pre-storage leukocyte reduction.
The NIH Clinical Center Department of Hematology has been performing some visual, morphologic and physical evaluation of the particles. They had received two products, one with what I considered type 1 but was adherent to the bag surface, and type 2 which was actually on a leukocyte-reduced bag from the national region that had a great number of floating type particles in it. The fibrinous material that was adherent to the bag was very similar to what other people have found that contained platelets, white blood cells, occasional red blood cells in a fibrin-like base. The second product that was post-leukocyte reduction, when those particles were separated and warmed in a saline suspension, the saline became milky, and they sent them off to the clinical chemistry service who determined that they were triglycerides without any cholesterol. So, they are going to be looking at some additional samples on that.
They have also done some paraffin blocks and those results are pending. Because of the chylomicrons that they experienced in the second sample that we gave them, they are going to do additional analysis on other products for us.
Dr. Moroff, at Holland Labs, is also going to be looking at donor characteristics. I didn't want to be very specific on any of their protocols because they are their own property and they are yet to be under way. They are going to evaluate products for either hard spins or soft spins. They are going to do the visual, morphologic, physical examination of the particulates in the blood. They are also going to look into the effects of pre-storage leukocyte reduction on white particulate material.
In summary, laboratory studies are under way on a number of fronts. Preliminary studies support previous reports of normal biological material, and results from additional studies are pending. Thank you.
DR. NELSON: Thank you. Questions? Yes?
DR. FITZPATRICK: Betsy, what is the timeline for these studies to be finished?
MS. POINDEXTER: Dr. Moroff, last week, had to go to IRB yet. For Dr. Stroncek it was going to be a couple of weeks before they could get under way due to time constraints within the blood centers. Everybody was under constraints. Because of the bad weather, and January being a low donation month anyway, they were fighting for every good blood sample they could get but it will be a couple of weeks at least. I think that the start to finish, probably two to four weeks, six weeks max. I think when we spoke earlier on the phone we had thought that it would be within the next three weeks. That was before we knew what their actual timelines were.
DR. NELSON: Previously we have talked about whether leukocyte reduction and filtration should be done at the time of collection or later. It looks like with this problem maybe it should be done both places, one to reduce leukocytes; the other to reduce the white particulate matter.
MS. POINDEXTER: We got a number of units from one region that had been leukocyte reduced, pre-storage leukocyte reduced and they had the dandruff or--
DR. NELSON: That is why I said that maybe we need a second filtration.
MS. POINDEXTER: Right. If I can go back to the gentleman who talked about Crispy Cream, if you note that these particles were first seen just after the Super Bowl--
--and being in the region, and Florida was playing in that game, you can just picture a lot of households with barbecues and fried chicken and fatback greens and everything else.
DR. NELSON: You know, we have identified the culprit and he is us. Dr. Binion?
DR. BINION: I am Steve Binion, from Baxter. Thanks again for the opportunity to come to the podium.
I want to start by just sort of cutting to the bottom line and let you know that at this point we have no evidence of any association, of any problems with Baxter BBPUs or those of other manufacturers associated with the observations of white particulate matter.
I will try and not duplicate any of the other presentations but just quickly, from Baxter's perspective, we were contacted on January 28th by ARC to inform us of reports of white substance, as it was termed then, that had been observed during component processing in Atlanta.
On January 30th there was again an additional update from ARC indicating that it had been noted in whole blood leukoreduced red cells and plasma. Our information at that point was that 67 collected products had been quarantined.
January 31st, we reported in a teleconference with FDA, the American Red Cross and CDC preliminary results from analysis of samples that had been obtained from Atlanta Red Cross, indicating that the material appeared to be composed of blood-related substances and, at the same time, we initiated a review of aspects of the BPU manufacturing while analysis of the samples continued.
February 3rd, as Dr. Page had reported, our scientists met in Atlanta with ARC staff to examine product samples and review the observed phenomena. At that time 21 samples were provided to Baxter by American Red Cross. These were made up of 20 samples of collected blood products and one unused blood pack unit.
From the company's perspective, we focused on five areas of investigation: review of the BPU manufacturing process; analysis of the phenomena that had been observed in ARC samples, and I want to stress that at the outset of this situation we invited both American Red Cross and FDA scientists to join us in northern Illinois on site at Baxter, and I think it was a very positive aspect of the collaboration and interaction with Red Cross that we did have two Red Cross scientists join us for the week of February 3rd and, from the FDA perspective, we also enjoyed meeting the investigators.
In addition to analyzing the particulate phenomena that had been reported, we also undertook chemical analysis of the anticoagulant and additive solutions, as well as chemical analysis of the BPU containers. The final area that we initiated studies involving the white particulate matter issue was focused on examination of blood component processing. Now, I am speaking in the past tense but I want to stress that our efforts to understand this situation are ongoing.
Looking quickly at the review of our BPU manufacturing process, we focused on product surveillance records, the manufacturing process itself and a review of batch records. From the product surveillance standpoint, we looked at reports to us over the past two years. This represented product history in excess of 500,000 BPUs and we had received only two reports of observations that could be termed similar to those that were reported out of Atlanta Red Cross. The manufacturing process and batch records review is still ongoing, still looking for anything out of the ordinary but at this point review of more than 50 lots of BPUs produced has not indicated any problem that could be associated or potentially correlated with these phenomena.
With regard to the analysis of the phenomena as they were observed in the ARC samples, and I think initially reported in photographs that were published, we undertook this activity the week of February 3rd with Baxter particle science laboratory scientists as well as histologists, etc., etc., working with the ARC scientists. These five categories that you see here represent the Baxter documentation of the ARC reported phenomena which were posted to the AABB web site.
I just wanted to point out that the descriptions and categorizations were developed in collaboration between Baxter and ARC scientists who were in our facility that week. Very quickly, five categories of phenomena were identified in the samples that were returned to Baxter from American Red Cross, microaggregates which were described visually as blood debris, clumpy blood debris. Category two was microaggregates. Tracking bubbles is the description in the Baxter report. That was explained to me as microaggregates which actually were in contact with bubbles in the units. Number three was bubble phenomena. These were unusual looking reflective bubbles that, upon further analysis, turned out to be bubbles. "Starry night," described to be white particles at the blood sheeting interface, and there was also oiling phenomena which was described as semi-liquid droplets in the seam area of the bag.
The microaggregates, at least the microaggregates that were examined in the samples that we had, were shown to contain leukocytes, hemoglobin and fibrin. That is category one. Category two was similar to category one but less formed.
Item number five, the oiling phenomena, were determined, at least in samples that we looked at, to be composed of a diffuse meshwork of proteinaceous material. The bubble phenomena, upon further analysis, were consistent with a saturated ester containing material that might occur in blood. There was no evidence of hydrocarbon or a silicone oil moiety present. Category four, the "starry night," was consistent with proteinaceous material and/or triglyceride or lipid material.
I think upon further analysis, our conclusion, based on the samples that we analyzed, was that the white particulate matter and other phenomena, at least in the samples that we analyzed, were consistent with blood-derived material.
As I mentioned earlier, one of the samples that was provided by ARC was an unused Baxter BPU which clearly, visible to the naked eye, had something that looked like white particulate matter associated with one of the bags in the container. Upon further examination, that material was determined to be on the external surface of the bag and it was shown chemically to be identical in composition to the white donor tubing segment pigment.
We tested file samples of several lots of the BPUs that had been implicated in the observations. What we found uniformly was that, as expected, there was an extremely low particulate burden in the solutions. There was no evidence of precipitation or any sort of suspension, resuspension anomalies with the solutions. We repeated on those file samples the chemical release testing that had been performed when the lot was originally manufactured and released to the field, and the product was still within specification.
Our next step in terms of chemical analysis was to examine the possibility that there was some difference in the blood pack unit plastic material's composition, or some change in that that could be correlated to the phenomena. What we found there was that all the products, both test and control--we tested what we call test products, BPUs from lots that had been associated at that time with the phenomena, and control products at that point were pulled from similar products that were not implicated in these observations. The bottom line there is that all products demonstrated chemical constituents related to the bag manufacture at expected levels, and testing to date indicates no evidence of any unexpected constituents in the bags themselves.
Finally, the last area that we looked into was examination of blood component processing conditions. I am not going to go through this slide in detail. I simply want to point out that using BPUs from Baxter as well as two other manufacturers, and you know who you are, we were able to replicate the majority of the phenomena that had been reported and observed in the ARC samples. Under the same conditions under which the phenomena were reproduced--again, I am not talking about all of the, you know, type 1, 2, 3, 4 but some combination of those.
In the instance in which the product was buffy coat reduced component, we were not able to replicate the white particulate matter finding. This is certainly not all-inclusive or all-encompassing, but it did confirm the possibility of reproducibly recreating these phenomena in bags from three different manufacturers.
Where we are at this point is that the company has decided to provide a grant to a group of blood banking and transfusion medicine experts. The hope here obviously is that this group will be able to study the situation and advance in awareness of factors that may have contributed to these recent observations of the aggregation phenomena in collected blood.
The panel will include Drs. Jeffrey McCullough, Roger Dodd, Ronald Gilcher, Scott Murphy and Merlin Sayers. At this point, as I said, we have not closed out our interest, nor our investigation into these phenomena. We look forward to the expert panel as it begins its activity. I want to just close by thanking you for the opportunity to give you an update on what Baxter has done in relation to the white particulate matter issue. Thanks.
DR. NELSON: Thank you, Dr. Binion. Any comments or any questions? Yes?
DR. FITZPATRICK: It is first I have heard that anybody has been able to reproduce the phenomenon.
DR. BINION: Okay.
DR. FITZPATRICK: Are you at liberty to share with us how you did that?
DR. BINION: Let me back up a little bit. First of all, I want to say that the processing conditions were developed based on discussions between our physiologists and ARC staff. Again, I want to stress that I think from day one in this situation we have been working in a very collaborative, very open manner with ARC, FDA and CDC as well.
Having said that, with the level of detail I have in front of me, I am happy to share. None of the units were leukoreduced. They were processed on the day of collection. They were cooled one and a half hours before processing. They were centrifuged at 4,200 rpm for 7.5 minutes at 4 degrees C and they were then stored refrigerated.
The ability to reproduce the phenomenon in the processed units in this instance appeared to be related to the handling of the components during processing and I believe the centrifugation aspects. Needless to say, we are continuing to look into this.
DR. NELSON: This reproduction suggests it may be the bag and not the donor. But I wonder if anybody has looked at donors who provided the potential transfusion recipients with white particulate matter, and whether or not the donor, on repeat donation, also did it again.
DR. BINION: The data that I presented here were obviously a laboratory study. We have those same questions. I think it is much more along the lines of what was described as Dr. Moroff's planned studies to look at the possibility of donor biological variability or contributing factors associated with the donation. Our efforts at this point were focused on understanding it and looking at whether there were conditions using bags from Baxter, as well as other manufacturers, that could shed some light on this situation.
DR. PAGE: I should point out that this has only been observed from whole blood donations and there is a required 56-day interval between--
DR. NELSON: Right, I understand that.
DR. PAGE: There was one autologous donor--white particulate matter was seen in several autologous units by us and I think, as mentioned, by someone else. One of those autologous donors who had white particulate material on one autologous donation donated again and it was not seen on the subsequent donation a week or two weeks later. That is an N of one.
DR. NELSON: Right.
DR. GOLDING: The data that Steve presented led us to believe, at least some of us at the FDA, that the interpretation is that the hard spin--the type of processing with the hard spin is a definite factor, and what could be happening here is first you take the whole blood and you put it in the cold and you activate the platelets. Then you do a hard spin and you remove the plasma so you are leaving the platelets, those activated platelets with the red cells in the bag, and that was generating particles, and those particles would be associated with platelets.
Now, my plea is, and I tried to say this and I probably didn't say it right after Dr. Page spoke, that I think we need to take all the data and look at it in terms of the type 1 particles which I think the evidence shows are the platelet particles. Taking all the data and looking at it in terms of all the types of particles is confounding and is probably imprecise.
I would like to ask, Dr. Binion, I mean, you did this, you reproduced it. On one of your slides you said all types of particles were reproduced. But the critical particles are the type 1 particles or the ones that are associated with platelets. To what extent was that represented?
DR. BINION: I don't have that level of detail available to me here. The point to even taking that path in the first place was to just understand if there was something reproducible or potentially reproducible in the processing that might shed some light on this. But I will get that answer for you.
DR. GOLDING: I think the same thing applies to Dr. Page, whether it is a particular bag or a particular process, I think it is worthwhile going back and saying, well, with the type 1 particles, which bags, which process.
DR. BINION: I think the only conclusion from that table that I presented is that looking at the processing of the components has to be factored into any effort to understand this phenomena. I think that is the only conclusion that I would draw from the information that I presented.
DR. KUEHNERT: Matt Kuehnert, CDC. Just a real quick question about the handling. You said you were able to reproduce it based on the handling. Could you elaborate? I mean, did you throw it out a window or tap on it?
DR. BINION: Actually, I didn't perform the studies. I believe the difference would have been gentle versus less than gentle handling. I think as this information is shared with the expert panel, and I think as reports from ARC and, hopefully, other studies are also made available to the panel that information would be forthcoming.
DR. ORTON: Steve, I will be able to give some more information about that in my presentation.
DR. PAGE: Now that I may understand Dr. Golding's question a little bit better, we have a spreadsheet that describes the number of units of each type from each region and correlates that with bag type and product code. That was emailed to the FDA about three weeks ago. So, the data is available that we can look at in that regard with you, which we would be glad to do.
DR. NELSON: Next is Sharyn Orton, FDA.
Follow-up ADR Monitoring
DR. ORTON: As you can tell, I have been dying to get up here.
What I am going to describe is some clinical studies that we are proposing and actually have ongoing right now at a couple of blood centers.
One of the things that the task force talked about was the safety issue. Even if we determine whether we can or cannot make these, or whether they are normal or not normal, we are concerned about the transfusion of these blood products. So, although adverse events have not been linked to transfusion of blood containing the recently described particulate matter, controlled studies are needed to further clarify the safety of such units.
In fact, there are reports in the literature that some particulate matter results from routine preparation of red blood cells, primarily by the hard spin method. The reports in the literature actually give 17-93 microns at the beginning of storage. After 21 days they go up to 9 g. There is quite a bit of literature from the '70s and the '80s that does describe this. So, it may be reasonable to assume that at least a certain proportion of these particles are, in fact, normal and a lot of reports I have gotten from my blood center colleagues who I have talked about these clinical studies do, in fact, support that type 1 particles are fairly normally seen and, when it comes to reproduction, can be reproduced fairly easily.
We know the particulate matter appears to be comprised of normal blood cell substances, seen less frequently in leukofiltered red cells, and in fact the leukofiltration appears to remove most of the observable particulate matter. We do have some preliminary data from ABC that would support that.
So, the scientific question is whether red cells with observable particulate matter can be transfused safely. Do these transfused red cells result in a higher than expected adverse event rate?
The other question we wanted to consider was does the presence of these pre-leukofiltration particulates affect the post-leukofiltration counts?
So, with the clinical studies we want to evaluate differences in adverse event rates between patients that receive units that were originally determined to have particulates and patients who received units that were not determined to initially have particulates. We also want to evaluate post-leukofiltration counts of these particulated red cells, both pre and post counts of white cells and platelets. Currently, we do have two blood centers that are participating in these studies.
What we know currently from these two blood centers is that, in fact, 60 red cells that very early on were observed with type 1 particulates were released for transfusion. The medical director at that center determined that they saw this all the time. Ninety percent of those units have been transfused. There have been no adverse events reported with any of those units.
As far as leukofiltration, one of the participating blood centers has looked at 32 red cells that had particulates. They leukofiltered them. The total comes to 76 between the two blood centers and, in fact, all of the post-leukofiltration white counts were below 5 X 106. In fact, over 70 percent were less than 1 X 106. One of the blood centers did do the pre and post platelet counts. They did 4, had the pre and post platelet counts done and the mean removal was 85 percent and the median removal of platelets was 92 percent. In fact, there is a reference in the literature from the early '80s that the leukofiltration filters do, in fact, remove about 80 percent of platelets.
What we have asked as far as adverse event rates, we have asked our two blood centers to perform a cohort study. When we consider the exposure is either the presence or absence of particulates in a prepared red cell, the outcome would be whether there is an adverse event. This can be done retrospectively or prospectively and our measure of effect is going to be relative risk.
What we are having the blood centers do is select particulated red cells, leukofilter them and from the same day and preferably hard spin also select two red cells that aren't observed to have particulates, that were also leukofiltered and that were shipped to the transfusion service so the transfusion service would get a unit that had particulates leukofiltered and also had two units that weren't particulated and were leukofiltered from the same day. The transfusion service will be provided with the whole blood numbers of these units for follow-up and the status of the type of red cell it was will be blinded so that they won't know whether we started with particulates or not.
We are asking the transfusion service to document either a report or lack or report of any adverse events for all of the whole blood numbers that are provided and, if necessary, provide some historic transfusion reaction rates just for red cells. We do have some information that suggests that for leukofiltered red cells the transfusion reaction rate in general runs around 0.1 percent, and for the non-leukofiltered red cells approximately 0.5 percent. The data is to be de-identified and sent to me for analysis. I am looking at both by center and preferably pooled data because of the sample sizes that are needed to actually get statistical significance with this kind of study. If necessary, further review or reported reactions might be warranted.
This breakdown is not correct. Currently we have 144 non-leukofiltered red cells that were observed to have particulate matter, determined to be type 1 and what is normally seen, and have been transfused or were released for transfusion; 136 leukofiltered red cells that had particulate matter have been released for transfusion. The corresponding essentially controls have already been identified and are currently in follow-up for adverse events. From 100 of the leukofiltered red cells that had the particulate matter, 50 have actually been transfused, with 94 of their control counterparts, and there have been no adverse events reported at all. Keep in mind that the sample size is not large enough to make a statement statistically, but this is where we are going.
As far as pre/post leukofiltration counts, we are interested in both the white cells and the platelets, as I mentioned. Again, data would be de-identified and sent to us for analysis.
FDA is asking blood centers to participate in these studies. It is very important that we get blood centers who can get the information from the transfusion services. We are not just interested in the blood centers that do 100 percent leukofiltration. We want any blood center regardless of what their percentage is. Depending on how we identify the different criteria, we are interested in all types of red cells, and the only way we are really going to get a large enough sample size, I believe, is to have larger participation.
What we have found from the blood centers that are participating right now is that by asking for reported adverse event data we are not asking for too much extra work to be done. Our original plan was to do chart review which we knew was not going to be practical when you start looking at hundreds, perhaps thousands of units. Anybody in any blood center that is interested in participating can contact me. My contact information is here. Thank you.
DR. NELSON: Thank you. Did you have any problems getting this through the IRB?
DR. ORTON: This data is all retrospective. You know, these were early on in the reporting when blood centers were starting to look at their units. Many medical directors determined that what they were seeing as far as the type 1s was that this was normal, and were leukofiltering because they did or were releasing them prior to us ever actually making a statement--
DR. NELSON: I understand that but you are proposing a prospective study and I just wondered if there were IRB issues with that because you wouldn't propose that the particulate matter would be beneficial. If anything, it could have no effect; it could be--
DR. ORTON: Correct, but our recommendation was also to take particulated units and leukofilter them and transfuse them. I mean, that is what our recommend is. Richard, do you want to make any comment about that?
DR. NELSON: I thought you were also transfusing leukofiltered units with particulates.
DR. ORTON: Well, as I mentioned, the study could also be retrospective. Some of the blood centers have tracked these kinds of things and there is some retrospective data.
DR. LEWIS: The two blood centers that Dr. Orton mentioned were identified after the units had been released for transfusion. They notified us that they had identified particulated matter in a certain number of units and had made their decision that they were still safe and released them. The FDA hasn't taken a regulatory stance. In all of our statements we are suggesting an enhanced inspection and quarantine but there is no guidance, there is no regulation that would prevent them from using these products.
DR. ORTON: In fact, as I said, there is no question that there were some particles that were determined not to be normal, but for others one of the medical directors determined that they see them on a regular basis and did not feel that these were abnormal; they see them all the time. Thank you.
DR. NELSON: Having tried to get a lot of projects through the IRB, I know that I would have had problems here and one would have to certainly get the informed consent, and that might be a problem if you did it prospectively. But I think the answer is important as to whether or not this has any adverse events, but it is a conundrum between getting an answer to an important question and dealing with the potential ethical issues as well. Dr. Kuehnert, from CDC, our final speaker today?
Centers for Disease Control
and Georgia State Division of Public Health
DR. KUEHNERT: Thanks for the invitation to do the final presentation at this BPAC.
I am speaking on behalf of the Georgia Department of Public Health who took the lead on this part of the investigation and we gave technical input. I think this is one of the rare studies I have been involved with where we were hoping to prove the null hypothesis.
First some background. These photos are courtesy of Chris Hillyer, type 1 matter.
I not even going to really read this. It has been presented just previously. Basically, after it was discovered that white particulate matter was found in blood products, the question was had any of it been transfused and, if so, whether there was an increase in reactions that were associated with transfusion.
So, our objective was to further investigate temporal association of adverse patient events with particulate matter. We queried transfusion services in Georgia to provide data on transfusion reaction rates over time; trends over time by component or supplier; and proportions of reaction types for one month, the month in which this was discovered.
We developed a one-page questionnaire to collect monthly data from January, 2002 through January, 2003, which I will refer to as the entire survey period. We asked transfusion services to report reactions as defined by their institution; the number of monthly transfusions by component type; the predominant supplier for the survey period; and, in addition, detailed information for reactions that occurred in January, 2003.
The detailed information for reactions in January included date of reaction; component type; component supplier, American Red Cross, community blood bank or Department of Defense; type of reaction, either hemolytic reaction, bacterial contamination, other febrile non-hemolytic allergic reactions or transfusion-related acute lung injury or otherwise or not specified.
A list of Georgia transfusion services was compiled using contact information from the American Red Cross, the FDA and the American Association of Blood Banks. The questionnaire was distributed on February 7th. A reminder was made to non-responders by the 11th, and date of collection was completed by the 21st.
Reaction rates were defined as number of reactions per 1,000 transfusions. Reactions were compared by component type supplier and time periods of interest. This was a little tricky in what to compare to because we didn't know when the problem actually started; we just knew when it was detected. So, we chose to use the month in which it was detected as the referent period and we chose just sort of arbitrarily the same month in the prior year and also compared it against January to December, 2002, or the remainder of the survey period. We compared proportions of reaction types within January, 2003 by week and half month, and used chi square and Fisher's exact test for statistics.
One hundred and thirty-one transfusion services were identified in Georgia, 7 couldn't be contacted or said they didn't transfuse anything over the survey period. So, there were 124 surveyed and 108, or 87 percent, of transfusion services responded, which I thought was really remarkable. In fact, we sent out the survey on a Friday at 4:00 and by Tuesday over half of the transfusion services had responded. I would say that by the end of that week about three-quarters had responded. So, I thought this was really amazing compared with what I have experienced before with surveys.
Over the survey period there were 1,213 transfusion reactions over half a million transfusions. That works out to about 2.3 reactions per 1,000 transfusions. The rates by products were RBCs, 2.42, you can see the range, 1.79 to 3.11; plasma, 1.48, a pretty wide range, 0.62 to 3.53. These are the rates for each month. These are the ranges for those. For platelets, 3.36, with a range from 2.12 to 5.42. Reaction rates did not differ when we compared January, 2003 versus January, 2002 overall, in other words, all components and all suppliers, or when we compared January, 2003 versus the rest of the survey period.
This just graphically displays that. You see the variability which was most remarkable I think in platelets and plasma compared to RBCs which was relatively flatter.
Next we looked by supplier, comparing January, 2003 versus the previous year and there was no significant difference. However, within supplier comparisons, when we compared from January to January, there were a couple of differences in just components when stratified and looking at community blood banks only. There was a significant increase in reaction rate associated with plasma, and a decrease in reaction rate, approaching significance, for platelets. You can see the numerator is very small, 0/6 and 6/1 but, nevertheless, that is what the statistics say. There were no differences in trends over the survey period for components collected by the Red Cross.
For community blood banks you can see there is a lot of variability in plasma, in this black line. The blue line is platelets, and a relatively flatter line for RBCs.
The numbers are much greater for Red Cross. I think it was over 80 percent of components that were transfused by Red Cross compared with community blood banks. So, the lines are flatter; the numbers are larger.
Looking at the detailed information in January, two were due to bacterial contamination. The majority were other febrile, noon-hemolytic reactions. There were some allergic reactions. Some were categorized as "other" and there were no TRALI cases reported. When we looked at the proportion of reactions categorized as an event type, there was no difference in the proportions when comparing by either half month or by week, sliced a couple of different ways.
Here it is just graphically displayed by component type. You can see by day the distribution over the month.
Lots of limitations to this. The goal here was to do it very quickly. We really wanted to make sure that it was just a one-page questionnaire, although when we got into the detailed data people had to sort of cut and paste it, but we really tried to keep it to one page. Of course, it was collected retrospectively so there is responder bias. We used the definitions of reactions by each institution so that introduces bias. If there was a sufficiently small adverse event rate due to units with particulate matter, we might not be able to detect it even with our large numbers, depending on what the rates were. Finally, there were changes in manufacturing--there could have been changes in manufacturing or modification of components over the time period that might confound potential trends due to particulate matter.
--we found in Georgia adverse monthly transfusion reaction rates of 2.42 for RBCs, 1.48 for plasma, and 3.36 for platelets per 1,000 transfusions over the survey period. For reactions reported in January, 2003 rates by blood component or supplier did not differ from the remainder of the survey period overall, and there were no significant changes in proportions of reaction types. There were some differences on stratified comparisons between individual months but those numbers are very small and there was a lot of variability during the survey period so our statisticians cautioned us about drawing conclusions. If we had used different time periods, such as quarters where there is sort of smoother data, we would not have seen those differences, but that is not what we chose at the outset so we were stuck with what we started with and that is what we found.
I would like to acknowledge Martha Iwamoto is the officer who did all the work, but couldn't be here today to present it. We had a lot of help from the FDA, especially in getting the transfusion service list. That was very helpful. We had Mary Chamberland who helped out a lot, as well as statistical help and some input from folks at Emory. Thanks.
DR. NELSON: Thank you. Well, it is reassuring data even though it is retrospective. Does this compare pretty much with what is in the literature, or what one would expect from the overall rates of transfusion by product? Is that pretty comparable with what other places have published?
DR. KUEHNERT: I would be curious just to hear what literature there is out there on it on a multi-center basis. I did a little bit of a literature search and I know that Dr. Linden has done a lot of work with transfusion errors and New York State has pretty comprehensive reporting, but I wasn't able to find published reports on transfusion reaction rates in a state or in a large number of centers.
DR. NELSON: Yes?
DR. ALLEN: Thank you. That was very different from the other presentations and I think rounded out the presentations nicely. Just a comment and a quick question, I am delighted to see actually that you did do it all retrospectively because I think had you tried to get retrospective comparison information with prospectively collected data, the difference in methods would have totally biased the period of prospective interest.
Question, information obviously went out to the transfusion service directors, and I suspect the Red Cross was in touch with them because of the various issues. Was there any information in the general media? Was this something that got picked up? Just how much was out there and what impact do you think that had on what happened?
DR. KUEHNERT: It is hard to say. You know, we set the cut-off point at January 31st. I think it hit the media over the weekend. There may have been some rumors before that where some transfusion services heard about it. You know, I think in looking at the reaction types there may have been some bias because we did notice that there were an awful lot of "others" and when we looked we actually had to weed out some because it said, you know, suspected something but then when we looked again it wasn't related to transfusion. We thought it was kind of odd to report that it is not associated with transfusion. Why would you report it? But overall we didn't see a jump in rate in January. So, maybe in the categorization it might have made a difference but otherwise, since they have a standardized way to report transfusion reactions, I don't think it had much of effect. We told them to just use the records they had.
DR. NELSON: Dr. Lewis is going to summarize this issue for us.
DR. LEWIS: I have some more slides in that previous side set. While you are loading that, I think that it is obvious from what we heard this afternoon that this was an incredibly intensive effort and required a lot of resources from a lot of different people, with incredible cooperation between regulatory bodies, other scientific organizations, as well as industry.
DR. NELSON: Did you tap into the bioterrorism research fund for this?
DR. LEWIS: Actually, the FDA Emergency Operations Center, which was organized in response to September 11, was activated and was used.
DR. NELSON: You have to find some way to spend this money I guess.
DR. LEWIS: I wanted to use the Commissioner's statement that was published on February 27th as a summary and conclusion, and it is really where FDA's thinking is right now.
Dr. McClellan issued an update on our investigation of the particulate matter and, as part of the title to emphasize it, there was not, and still is not, evidence for any increase in adverse events associated with the particulate matter.
The exact cause is still unexplained. There is no evidence of a threat to blood safety, although there has been all along some concern about blood supply.
The investigations at a number of different places showed that the particles themselves are composed of normal blood substances, primarily platelets, and the rate of adverse reactions has not increased. This is supported by the studies that you heard about from Dr. Kuehnert and Dr. Orton.
All analyses have been negative for infectious agents, chemical contaminants and any defects in the blood bags.
In spite of initial reports, the particulates were not limited to the Red Cross collections or to Baxter blood bags.
Again emphasizing, no evidence of adverse reactions that are associated with it.
Baxter's testing showed no changes or obvious differences in process in particle composition or any of the materials that they use.
This is the short list of participating individuals and organization.
This goes to the predominant current theory of the cause.
This has to do with how blood is processed today. When whole blood is hard spun in order to collect plasma, and if there is not to be a platelet collection from the particular units, this hard spin spins the platelets and white cells into the red cells. The platelet-poor plasma is collected and the particulates have been predominantly observed in these types of red cell units. When platelets are to be collected, the whole blood units are soft spun and platelet-rich plasma is expressed. Currently the use of apheresis platelets is much more predominant today than it has been in the past. Thus, there are more units in the last decade that would use the hard spin to collect red cell units. As a result of that, the observation of particulated units in red cells may be more frequent.
We currently think that there is still some uncertainty with what is causing this. We encourage people to continue with their enhanced visual inspection, distinct from the required visual inspection. It may not be necessary--if the units are leukoreduced, to go ahead and use units that have been leukoreduced even if particulate were observed before leukoreduction. Again, the enhanced visual inspection seems to be beneficial only before units are leukoreduced.
In centers where it is not normally leukoreduced and particles are not observed, it is probably safe to assume that the units do not pose an increased risk.
Let me just say that the work ahead will at least require accumulation of a lot of data and, hopefully, we will be able to draw some conclusions from all of those data, if nothing else, to define what has caused these particulates and possibly use that information to improve the quality of blood.
I would like to also take the opportunity to thank all the presenters who took the time and effort to put together presentations for us today, all of whom did this on short notice and presented a lot of information. Thank you, Dr. Nelson.
DR. NELSON: Thank you. Yes?
DR. FITZPATRICK: Richard, I want to say one thing. I think the FDA's initial response was very admirable to what occurred and the way they approached it, but I think now we seem to be mired in the need to explain the phenomenon and devote a lot of resources to it when, at the same time, we were briefed on a non-plan on how to deal with West Nile Virus.
So, I am concerned about the priorities here. This is a very interesting scientific phenomenon for a lot of people; they want to devote a lot of time and resources to it. Yet, in a practical sense there appear to be no adverse events. It is 12 million donations a year from 12 million different biological individuals and the proportion of observation is about the same as the manufacturers' market share. Now, that we have had our peak of intense scrutiny we seem to not be seeing it anymore, maybe because people don't want to deal with it anymore. It is hard to tell.
So, I question the rationale behind the need to explain something that may never get explained, in a manner that does that which may be taking resources away from everything else. There needs to be an "extraction" policy as well as a "dealing with" it policy.
DR. LEWIS: Thank you very much. I think no one appreciates your comments more than those of us at the FDA who have been dealing with this. We recognize the need to draw some important conclusions from all of these efforts and we would like to do it as quickly as we can.
DR. NELSON: Dr. Smallwood is going to announce the next meeting where we will get a further update on white particulate matter.
DR. SMALLWOOD: Just to let you know, the tentative dates that are set for the next Blood Products Advisory Committee meeting are June 19th and 20th. It will be in this hotel. So, please use your usual resources for confirming this and tell your friends that left. Thank you.
[Whereupon, at 4:45 p.m. the proceedings were adjourned]