FDA Home Page | CBER A-Z Index | CBER Search | Contact CBER | CBER Home Page

DEPARTMENT OF HEALTH AND HUMAN SERVICES FOOD AND DRUG ADMINISTRATION CENTER FOR BIOLOGICS EVALUATION AND RESEARCH SAFETY AND EFFICACY OF METHODS FOR REDUCING PATHOGENS IN CELLULAR BLOOD PRODUCTS USED IN TRANSFUSION VOLUME I Wednesday, August 7, 2002 8:02 a.m. Jack Masur Auditorium Building 10, Clinical Center National Institutes of Health Bethesda, Maryland 20892 P A R T I C I P A N T S Speakers and Moderators Jaroslav Vostal, M.D., Ph.D., Chair James AuBuchon, M.D. Celso Bianco, M.D. Mo Blajchman, M.D. Mark E. Brecher, M.D. Michael P. Busch, M.D., Ph.D. Richard D. Diamond, M.D., MPA Roger Y. Dodd, Ph.D. Sunny Dzik Jay S. Epstein, M.D. Mahmood Farshid, Ph.D. Hanan Ghantous, Ph.D., DABT Mindy Goldman, M.D. Margarethe Heiden, M.D. John Hess Sukza Hwangbo Harvey Klein, M.D. Matthew Kuehnert, M.D. David A. Leiby, Ph.D. Gary Moroff, Ph.D. Albert E. Munson, M.D. Scott Murphy, M.D. Paul M. Ness, M.D. Anita M. O'Connor, Ph.D. Betsy Poindexter Toby L. Simon, M.D. Sherrill J. Slichter, M.D. Edward Snyder, M.D. Suzanne R. Thorton, Ph.D. Steven Wagner Mark Weinstein, Ph.D. Dr. Hannelore Willkommen Roslyn Yomtovian, M.D. C O N T E N T S Evaluation of efficacy of decontamination methods Welcome - Jaroslav Vostal Introduction - Jay Epstein Session I: Transfusion-Transmitted Pathogens Bacteria - Roslyn Yomtovian Viruses - Mike Busch Parasites and other pathogens - David Leiby Clinical consequences of transfusing contaminated blood products - Richard Diamond Session II: Pathogen reduction methods in blood components Overview of different methodologies - Steve Wagner Evaluation of viral decontamination - Panel Discussion CBER Evaluation of methods for viraldecontamination for non-cellular transfusion products - Mahmood Farshid European Experience in decontamination methods evaluation - Hannelore Willkommen European regulation of blood component's quality and safety - Margarethe Heiden Moderator: Mike Busch Panel Discussion - Steve Wagner, Mahmood Farshid, Hannelore Willkommen, Margarethe Heiden, Roger Dodd, Harvey Alter Lunch Evaluation of bacterial and other pathogens decontamination Moderator - Mo Blajchman CBER list of bacteria for testing decontamination methods - Betsy Poindexter Bacteria identified in BaCon study - Matt Kuehnert Bacteria identified in Hopkins study - Paul Ness Canadian experience with bacterial contamination - Mindy Goldman Panel Discussion - Mindy Goldman, Roslyn Yomtovian, Mark Brecher, Paul Ness, David Leiby, Matt Kuehnert Industry representatives - presentation of data in Phase I/II Moderator - Celso Bianco Baxter - Larry Corash Vitex - Bernadette Alford Gambro - Ray Goodrich Question and Answer Session FDA perspective on Day 1 and Close - Dr. Vostal P R O C E E D I N G S DR. VOSTAL: Good morning. My name is Jaro Vostal, and I'd like to welcome you to Washington, D.C., NIH, and to the workshop on pathogen reduction in cellular blood components. We have a very exciting day today, packed with lots of information and discussion, so we have to get started. We will also try to stay on time because it's important for us to cover all the topics from morning until afternoon. There's one order of business I'd like to bring up, and that is, people have approached me about whether we've been collecting information on conflict of interest for the speakers. Since this is a public workshop, we don't have a policy to do that, only for advisory committees. But if you're a speaker and you have some potential conflicts, we welcome you to reveal those on a volunteer basis at the beginning of your talk. So, to get started, we have Dr. Jay Epstein, who's the Director of Office of Blood Research and Review, and he will give the introduction this morning. DR. EPSTEIN: Well, thank you very much, Jaro. It's my pleasant task to set the stage, but I think it's obvious that the real thanks go to Jaro and his group for developing and organizing the program, and special thanks also to Joe Wilczek for handling our logistics. As Jaro said, our goal is to review the emerging technologies that are applicable to pathogen reduction for cellular blood products. And how do I get the first slide on? Just with some acknowledgment to the cartoonist, I think it's obvious to everyone in the room that bacterial, viral, and protozoa pathogens have been identified in blood products, that mortality and morbidity are associated with transfusions that are contaminated by these pathogens. FDA and industry as a whole are committed to reducing the incidence of pathogen contamination in blood products. And as you can see from the organization of the program, we are mindful that there are a number of different ways to approach the challenge of decreasing contamination. Broadly speaking, these methods include efforts to primary prevent contamination of the collection. This includes the skin preparation method as well as the possible effectiveness of diverting an initial volume of the collected blood away from the final storage container. We then move to efforts to detect pathogens in the collection as a way to avoid use of contaminated units. And we have, again, a set of technologies directed at different pathogens: nucleic acid testing, bacterial culture, use of bacterial fluorescent probes, and some novel biochemical tests for pathogen proliferation. Failing that, there's then the new technology challenge of methodologies to decontaminate a contaminated unit, and we have targets that are directed toward nucleic acid, and these include both methods that are chemically spontaneously reactive as well as those that require activation by UV light. So what we hope to do at this workshop is to evaluate these several different approaches to pathogen reduction. From the FDA's standpoint, each individual approach will need to be evaluated both for safety and efficacy within its context of intended use. Of these three basic methods, the use of decontamination is both the most novel and the most complicated in terms of the safety and efficacy assessment and, therefore, will merit a considerable amount of our time. Decontamination methods involve the addition of mutagenic and potentially carcinogenic chemicals. We are aware that residual amounts of these chemicals can remain within the transfusion product and then be transfused. And we also recognize that in some cases these chemicals may interact with the product itself, potentially changing its character. The focus of the workshop, again, as noted earlier, is on the cellular product, and the cellular products, as we know, are unique because they are, for the most part, not frozen. And storage temperatures above freezing do allow particularly for bacterial proliferation. Additionally, these cells are susceptible to damage both from chemical exposure and UV light treatment, and the evaluation of toxicity or damage to the cells is itself a complex task. The intent, then, at the workshop is to promote the discussion of the scientific aspects for evaluating pathogen reduction in cellular products. We hope to hear public opinion on the appropriateness of the approach that the FDA is taking toward evaluating these methods, and we hope to encourage the development of novelty contamination methods by outlining the necessary steps toward validating clinical use. We then will review the different approaches to evaluating pathogen decontamination methods. We hope to establish the appropriate methodology for testing efficacy, and we hope to obtain some level of a scientific consensus on the minimum level of efficacy that will be required. Concomitantly, we hope to get a framework in place to evaluate toxicity of these methods. We will then review the current methods for evaluating the efficacy and the safety of cellular products. And we will specifically focus on the question of FDA's approach toward assessing toxicity when there's a question of mutagenicity or carcinogenicity. So, as you can see, we've set ourselves a rather ambitious task, but we know that we have assembled in the room some of the leading experts in these areas, and we look forward very much to the insights that we may gain from this dialogue toward establishing the framework of FDA decisionmaking that will permit product development to go forward. So I thank you all for coming to the meeting and look forward to your contributions throughout the sessions. Thank you. I give the podium back to Dr. Vostal. [Applause.] DR. VOSTAL: Okay. Thank you, Dr. Epstein. So now we'll move into the first session of the workshop, and this will deal with an overview of the different types of pathogens that can be transmitted by blood transfusion, and our first speaker is Dr. Roslyn Yomtovian, who is the Director of Blood Bank, Transfusion Medicine Service and the Acting Director of Clinical Pathology at the University Hospital of Cleveland, Ohio. She will cover the bacterial contaminants. DR. YOMTOVIAN: Well, thank you so much, Dr. Vostal, and others, for inviting me back again. I feel like I'm in the movie "Same Time, Next Year," for those of you who are familiar with that movie. But I'm certainly happy on behalf of our group in Cleveland to talk to you about this very important topic and give you an overview on what I feel are the key issues in bacterial contamination. To start on a little light note here, there have been two other workshops on a related topic in the past, so the question is how many acts will we finally have. I hope that this is indeed the last act before we finally do something about what I consider a very important issue in blood transfusion safety, as I hope to remind you or convince you of before I'm done with my talk. And so, really, I'm going to cover with you very, very briefly two key issues, and they are to recognize the clinical significance especially of platelet bacterial contamination, and then very briefly pick up where Dr. Epstein left off, just going through what are some of the strategy or strategies to prevent or interdict the problem. Now, I will focus on platelets, and in trying to convince you that that's appropriate, I paraphrase Willie Sutton, who actually never said that he robs banks because that's where the money is. He said--"I rob banks because that's where the money is" is not really what he said. It's modified from the irony of using a bank robber's maxim as an instrument for teaching medicine is compounded, I will now confess, by the fact that I never, never said it. Why did I rob banks? To me the money was the chips, that's all. So, to me, platelets are the chips in terms of bacterial contamination, that's all. Actually, Willie Sutton wrote two books, believe it or not, so he did more than just rob banks. Well, to begin with, a little bit of an overview on the subject, and I don't think I have to convince anyone here of what's written on this slide: In an era in which the risk of transmission of recognized transfusion-transmitted viruses, particularly HIV, has been virtually eliminated, it is paradoxical and somewhat ironic that the earliest recognized infectious transfusion complication, bacterial contamination, is now the most frequent and indeed is the most daunting and proving to be the most difficult to eradicate. And we fell into this not because, I must say, I was born with a genetic inclination to study this, but because about ten years ago we had a cluster of four episodes of bacterial contamination at our facility, which was thoroughly investigated by the CDC and FDA and was reported in MMWR and later expanded into a fuller report in Infection Control and Hospital Epidemiology. And, in essence, what we discovered, what everyone discovered was we weren't doing anything wrong. That was our great fear in the beginning. What was unique about our facility? Why did we have all these contaminations? We were doing nothing wrong. Our technique was proper, et cetera. Since then, of course, we discovered that this is a problem, for whatever reason--and I don't have time to go into the details--that does seem to cluster. It's a very interesting aspect of this problem, and perhaps we can talk about it at the panel. But I want to now just take a couple of minutes and give a very truncated, historical overview, really starting with the platelet story. Of course, contamination of blood goes way back to the early part of the century, if one looks at red cells. But I want to look at platelets, and, therefore, I'll go back to 1969 in a very important paper by Murphy and Gardner in which they discovered that you could increase the shelf life of platelets greatly by storing them at room temperature. And, of course, development in plastic bags made that possible. But one of the things they worried about was if there would be a risk from inadvertent bacterial contamination. Now, they felt there was not a risk, and that was based on a very small study--very small. I think "n" was, you know, no more than about 100 in their study, and, therefore, it was not surprising that only a couple of years later Buchholz's group reported on a case of transfusion-induced enterobacter sepsis, and they admonished that, although there is utility to room temperature storage of platelets because you could keep them longer, the platelets are functional longer, there may be a serious risk to those receiving such products. Platelets stored at room temperature should be used with caution, especially in high-risk populations. And shortly thereafter, they used that incident as, of course, we've used our experience in this area to do a larger study, and they found that up to 1.6 percent of platelet units are contaminated, and they noted that to be storage-time-related, as many others have noted since. And they further said that the risk of bacterial proliferation may warrant a review of current methods of platelet collection and of ambient temperature platelet storage. And, again, they cautioned use of these products, especially in recipients of impaired host defense mechanisms. And, in retrospect, it fell largely on--it fell silent. There was really not an audience for what Buchholz and others were saying. Another group--this is not the same Dr. Jacobs in our group--at about the same time, again, worried that platelet concentrate stored for four days at room temperature would facilitate bacterial proliferation, and in a very prophetic recommendation, which I don't know if they ever followed up on, but it certainly rings true for today, they proposed use of a direct film made from a sealed segment of the tubing incubated at 37 degrees overnight when the parent bag is stored at 22 degrees, and they said that should provide a reliable indication of bacterial contamination at the time the platelet concentrate is being distributed. That's really the essence of what's going on now with the various culture schemes where you have a holding period at a higher temperature to encourage or augment the bacterial growth. Very prophetic. Well, the saga went on. New bags were developed to store platelets, and instead of being worried about bacterial contamination, the storage time of platelets was actually extended with the new generation of bags from five to seven days, and it was only then in 1986, in response to an increase in number of reports of platelet transfusion-associated sepsis, that the Blood Product Advisory Committee to the FDA recommended going back from seven days to five days. And, actually, if you read that report carefully, you could make a case that they should have gone back to four days. But I think this was a compromise at the time of trying to keep the supply up and reduce the problem. So what is the risk of contamination in platelets? If one looks at the FDA's own data derived from the mandatory reporting of transfusion-associated deaths, one notes that in the two time frames, an earlier and a more recent one, although the total number of deaths is about the same, the percentage due to contamination has gone up. And that's likely due to the increased use of platelets in this time period, which is greatly increased, and perhaps some better recognition. Oops, sorry. If one--I am going, I think, the wrong way. No. Sorry. Okay. I'm not going to belabor the BaCon study because you'll hear about that later, but that is the second way that the risk has been evaluated of bacterial contamination. And, of course, that study did use rigorously defined criteria to capture cases. And I'll get back to that in a little bit because, by being so rigorous, obviously the total number of cases would be limited. And, in fact, in the BaCon study published results--and, again, I'm sure you'll hear an update of this later--here are the results. There were five cases in red cells and there were 11 cases in pooled random units and 18 cases in single-donor units with fatalities in all of these cases, a much higher rate in the platelets. Not to say the problem doesn't exist in red cells, but it's certainly far greater in platelets. Now, both the FDA reporting mechanism and the BaCon study are dealing with the tip of the iceberg, and I'll explain that shortly. The work we've done through the years is trying to look at the rest of the iceberg, which I will try to convince you there are many clinically significant cases in here. And even cases that are unlikely to be clinically significant may be important in an epidemiologic sense or to warn us that something may be percolating. So these cases, I believe, are also important to recognize. And so what I have done is tried to summarize on one slide--and this is probably the most important slide that I will show today--that if you normalize our experience to 100,000 platelet transfusions in the denominator, I'm comparing the risk of transfusion-transmitted bacterial disease per 100,000 transfusions and an estimate of transfusion-transmitted deaths from this problem in the BaCon study and in our experience. And the differences are, you know, orders of magnitude different. So BaCon would have one per 100,000 of bacterial transmission. Our number suggests it's something like 200 per 100,000. And deaths, BaCon reported 0.2, if you look at their publications, in transfusion per 100,000; in our experience, it's approximately 10 cases per 100,000. Very different numbers and a very significant problem. And work that we've done also showed that unit per unit, the risk of contamination in a single-donor platelet apheresis unit versus a random unit is statistically the same. Now, obviously, since random units are pooled, the risk is much greater, approximately--you multiply the risk by the number of units in the pool. But unit for unit, the risk appeared to be the same. I wanted to share with you just a few very concise clinical vignettes from our studies through the years on bacterial contamination of platelets as a way of illustrating why the data from a study like BaCon, which is a very useful study, but it's very restricted in the numbers of cases that were finally reported because of the rigorous criteria that it used, and also it was voluntary and it certainly wasn't a prospectively designed study, as were our studies for many years. So let me share with you a few cases. A 45-year-old patient was receiving multiple antibiotics, developed shock beginning 30 minutes after a transfusion of an apheresis platelet, Streptococcus bovis was isolated from the platelet bag, but her blood cultures were negative. So it wasn't accepted as a BaCon case because it required that the blood cultures be positive. But many patients that receive platelets are protected in the sense--protected from the growth of the organism, certainly not from the endotoxin, because they're on antibiotics. This patient did not die from this event but died shortly thereafter from an unrelated cause. Another case is a 63-year-old patient with AML who developed rigors 15 minutes after a transfusion. No other signs or symptoms. Was given a pool of five random platelets. A very astute nurse, however, obtained two blood cultures right about that time, and the platelet pool bag was part of our surveillance, prospective surveillance program, was positive for coagulase-negative staph, as were the two blood cultures. She was treated with vancomycin, recovered. This was accepted as a BaCon case because the organisms were identical by RFLP. The blood cultures were positive. The pool bag was positive and so forth. But I venture to say very few hospitals would do blood cultures on someone only having rigors after a platelet transfusion with no other signs or symptoms. Two more very interesting cases are a 27-year-old with ALL who received a pool of platelets, and I will say uneventfully, absolutely uneventfully. Twenty hours later he felt chilly, and 22 hours later after he got the transfusion, he spiked a very high fever. We were doing surveillance cultures on our bags. That was positive by then for Staph aureus. A blood culture was drawn. It was also positive for Staph aureus. He was treated with vancomycin and also required granulocytes. This was also not accepted as a BaCon case because of the delay in the symptoms, significant delay. And we've seen that more than once. A 72-year-old with aplastic anemia received five random platelets in a pool, again, I emphasize uneventfully. We were doing surveillance cultures of our platelets. It was positive after the fact for coagulase-negative staph. Two blood cultures were obtained at day four. One of the two cultures was positive for what seemed to be an identical organism, at least with antibiotic susceptibilities. A corresponding red cell unit from that donor was also positive. She was treated with vancomycin, and, again, that was not accepted as a BaCon case because there were no clinical symptoms. So the point of this is simply that the problem is far greater than what has been reported to the FDA and what has come out from the BaCon study. And this just summarizes what I've already mentioned. In the remaining time, I want to just very briefly and quickly review the different strategy or strategies to interdict bacterial contamination because it's possible that, unlike the paradigm in blood bank for many years in which you wanted, you know, one strategy to deal with the problem, we may need multiple strategies to deal with this problem, although I'm sure that will be a topic for discussion later. So I've just put these into what I consider four paradigms: the bacterial contamination avoidance methods, inhibition methods, detection, and elimination methods. And, of course, this meeting is going to focus primarily on the elimination methods. But starting at the beginning, an avoidance method would certainly be the ideal approach since it would avoid the need downstream for bacterial detection, growth inhibition, or elimination. And avoidance strategies depend on the bacterial source, obviously. There are two ways, two main ways that blood products can become contaminated. One would be donor bacteremia, and certainly with platelets, the most common way is donor phlebotomy. Of course, to deal with donor bacteremia, the ideal way, if possible, would be donor screening. Obviously that won't be 100 percent. And for phlebotomy, an ideal way would be to have the best arm preparation possible. But even with the best arm preparation possible, you're not going to eliminate this problem because it's been shown fairly convincingly that bacteria don't reside only on the superficial skin surface, but bacteria are harbored in the deeper layers of the skin and skin appendages. And, for example, the American Society for Microbiology even allows a blood culture contamination rate of up to 3 percent, and in part, that reflects the fact that you cannot totally decontaminate the skin. In a table that was published by Dr. Ernst in one of--a throwaway journal but a very useful table, nonetheless, on the percentage of organisms as contaminants, it's interesting that certainly two of the ones that we see commonly in platelet contamination, bacillus and coag-negative staph, are also the ones that are found most commonly as false positives, quote, false positives, in blood cultures. So certainly some of these are likely related to the phlebotomy process. Now, there is likely a correlation between the type of skin prep and the rate of culture positivity. Many papers through the years have noted that iodine tincture is a superior microbicide compared to povidone, iodine, and other methods. In fact, you in some papers get a 50-percent reduction in spurious contamination, and you even get a reduction in the quantitative level of bacterial growth. However, a very recent paper by Calfee, et al., in the Journal of Clinical Microbiology, actually refutes that and doesn't show a statistically significant difference in these methods. So I'm convinced there likely is a difference in methods, but I'm not expert enough to tell you what the best method is. It's certainly something that needs ongoing evaluation. Likely of much more importance is the training of the phlebotomists, and in three studies--two by Weinbaum and one by Schifman, et al.--they compared the rate of contamination of trained phlebotomists with non-phlebotomists, and this is very germane because in an era where hospitals are trying to cut costs every which way and practically take people off the street to do certain tasks, really this is food for thought, that people that don't know what they're doing will have a much higher rate of contamination than trained people. So this is a very important point. So, in summary, based on what has previously been published, it appears that iodine tincture disinfection is preferable; use of trained personnel is important. Obviously, the phlebotomy site must be selected with care. Scarred areas and sites near indwelling lines harbor more bacteria, so those need to be avoided. The phlebotomy site needs to be prepared with care. It's been shown that use of friction when prepping the site will disinfect it better, and, of course, one must allow the disinfectant to dry and not touch the site. And certainly if one uses single-donor platelets, you statistically reduce the risk of contamination because it's one venipuncture versus multiple venipuncture, so the rate will be lower. Now, switching to a very brief overview of bacterial growth inhibitory methods--and I'm really just touching on some highlights here. I've already mentioned that there's--and Dr. Epstein has mentioned, bacterial platelet contamination is linked to room temperature storage and the time of storage. If you reduce the storage time, you'll reduce the problem. It's not that the bacteria aren't there. They just need time to grow up to clinically significant numbers. So if you could refrigerate or freeze platelets, you would greatly reduce the problem. The problem is that we've already learned from Murphy's work et al. that cold temperature irreversibly damages platelets, so that's a problem. You'd need a cryoprotectant to protect the platelets at colder temperatures. And there has been a lot of work on this ongoing to figure out a method to preserve platelet function when they're stored either frozen or at 4 degrees. And the work that's most successful and is ongoing in this is work from Life Cell Corporation in which they have now some studies that have already been published and some studies that are in progress showing that when you use their proprietary agent called Thrombosol with a lower concentration of DMSO, and you use that to freeze platelets, it's a much easier method to employ. You don't need mechanical freezing, and you could thaw the platelets quickly. And, furthermore, you don't need to wash out the agent before transfusion, and you get equivalent or improved post-thaw platelet function. So although this is really not ready for clinical use at this point, there may be something here to keep following. So switching gears to bacterial detection methods, the challenge here is: What is the level of clinically significant or clinically tolerable bacteria that you want to detect? Will you allow a level of 101 or 102 and say I don't care about it, I'll only detect 103 or 104? What is that level? Well, based on our experience, I believe the level is around 102-3 CFU per mL, because from our experience through the years, we've shown that a level of even that range may be clinically significant. The ideal method would be rapid, inexpensive, sensitive, specific, practical, and simple. That's why I say there may not be one method. Testing should be as close to transfusion as possible to enhance detectability. If you test too early, the bacteria will not have grown up enough to detect them. So you need to test close to transfusion and to remind people you should never do the sampling from a segment or a link that's made at the time the original product is made. Statistically there will not be enough bacteria to be present in these small volumes that are in these to be positive. So this next slide summarizes--and I've adapted this from a nice review article that Dr. Brecher's group wrote a few years ago--the different approaches to bacterial detection. And the one group of approaches that's being used most successfully are under the category that I called "cell growth," so culture methods which are very sensitive and a type of surrogate culture method, percent oxygen in air, which is a method that's being developed by Pall Corporation. There are certainly other methods that I've put under cell marker methods that are based on antibiotic probes or antibody probes or a very new method that's coming on the scene based on epifluorescence microscopy, which may be quite sensitive, although maybe slightly less than cell growth. For some reason, the molecular biologic approaches haven't been as successful, but that may be a technological issue rather than an inherent problem with these methods. And this does not include all detection methods, but it's just designed to give an idea of order of magnitude of sensitivity of different methods. A Gram stain, which we did for many years, is extremely insensitive. You need about 106 CFU per mL to detect bacteria; whereas, culture methods are right down around 101, 102 CFU per mL and right now really are the most successful in preventing this problem in institutions that are using this method. This is a summary of our cases of contamination over not quite ten years, clinical outcome, number of isolates, the specific bacteria, and CFU per mL because we were doing quantitative cultures on the bag in all the implicated cases, and, again, these are largely based on prospective studies. The point is the red numbers here indicate positive blood cultures. So even numbers as low as 103, in one case as low as even 102, were linked to some sort of clinical symptom, even if they were only very mild symptoms. So I think the issue of what level you'd like with detection is, as I've indicated, ideally around 102-103. Whether that can be achieved by anything other than culture would remain to be seen. This just gives a little bit of our experience with Gram stain. We stopped doing Gram stain in 1999. We only ever interdicted--we did interdict some cases, so I'm not going to knock Gram stain completely. But, obviously, many cases we did not interdict using a Gram stain because it's very insensitive, and really, you're only going to get rid of the absolute--you know, the blood products that are literally cultures at that time. So, shifting gears to bacterial contamination prevention methods, filtration, phlebotomy diversion, and, finally, you'll hear a lot more about the photochemical decontamination have been examples of prevention methods. Early on filtration was shown to be reasonably effective for Yersinia enterocolitica. But, of course, that's a problem largely of red cells, and, of course, now that universal leukoreduction is sort of the trend of the future, this problem is likely to become less and less, and it certainly is not going to solve the problem with platelets. But it has a role, a limited role, especially with select organisms, and, for example, Yersinia. Regarding the diversion technology that Dr. Epstein mentioned, there have been several articles published on this. There have been two clinical type articles from Europe. This is an older article that was published in Vox Sanguinis that showed that in about 3,500 whole blood units that were collected over four months, two bags, two 15-mL bags were attached in series--maybe not in series, but in sequence. Blood was diverted first to P1 and then to P2. And what they showed was in the P1 bag, the blood that was diverted, the initial diversion was much more frequently contaminated than the P2 bag, suggesting that diversion may be effective in reducing the level of contamination. A very recent study appearing in this month's issue of Vox Sanguinis took 7,000 whole blood units which were collected with something they call the camposampler attached to the blood collection system to divert the first 10 mL, and then they tested the sampling bag by the BacT/ALERT following collection, and they had about 18,000 control whole blood cultures. Now, their results are provocative and interesting. They did a one-sided test, statistical correlation test to look at the reduction or the incidence of bacteria in the study versus the control group, and doing the one-sided test, there was kind of marginal statistical significance for overall reduction in bacteria with the diversion group versus the control group. But, very interestingly, using just standard, regular statistics, two-sided, they found that there was a very significant reduction in coagulase-negative staph in the study group versus the control group. So it may be that the diversion method may be particularly apropos and good for prevention of coagulase-negative staph. So I refer you to that study. I'm really not going to spend time talking about this because this is really what this meeting is about. But to date, there have been publications on two broad methods of photochemical decontamination. The psoralen method of decontamination, which forms an adduct with nucleic acid and has been shown to be quite effective, although not absolutely effective--but you'll heard about that, I'm sure, later--in eliminating bacteria. There is a little bit of inconsistency, and the second different method is based on a riboflavin B2, which is an oxidative process, and it also has been shown to be somewhat effective in reducing bacteria, and I'm sure we will hear that data at length so I'm not going to dwell on that. So, in summary, bacterial contamination is an ongoing, recurrent complication of primarily platelet transfusion therapy. There is no systematic approach in the U.S. to reduce or eliminate this problem at this time. We haven't even defined an approach, let alone implemented an approach. So I hope what comes out of this meeting will be some approach. A single ideal preventative strategy--safe, rapid, simple, inexpensive, sensitive, specific, and practical--has not been developed. In fact, it's unlikely that there will ever be one single such strategy. So we may be best to look at combined strategies, because there are numerous strategies alone, but better together, that not only will reduce--certainly alone they'll reduce, but perhaps together they may eliminate the occurrence of this problem. And, finally, and very importantly, prevention strategies for once may prove cost-enhancing if linked with a seven-day platelet storage product. And I know we'll be hearing more about that at this meeting, and you can't ignore costs in this day and age where everyone is under the gun always to save, to save money. And I just thank the members of our group who worked on this through the years because I'm only just one of the people that happens to come and speak at meetings. But Dr. Jacobs and a crew of very talented microbiologists, Dr. Sepatnekar now at the Cleveland Red Cross, Dr. Palavecino, our current blood bank fellow, and Sara Lee and Ann, who are very talented research microbiology techs, have certainly done much of the work through the years. I thank you for your attention. [Applause.] DR. VOSTAL: Thank you, Dr. Yomtovian. We have time for one burning question, but hopefully it will be short. Anybody have a question? [No response.] DR. VOSTAL: Okay. Thank you. We will now move into viruses, and this presentation will be given by Dr. Michael Busch, who's the Vice President of Research and Scientific Affairs at the Blood Centers of the Pacific, and also a professor of laboratory medicine at UCSF. Dr. Busch? DR. BUSCH: Thank you, Dr. Vostal. First, in terms of conflict, I think I've worked with every company that both tests and inactivates, but none of what I say, I think, will have any relevance to conflict. What I've been asked to do is to update on current risks of virus infections from transfusions, and then particularly discuss the issues of the levels of viremia in these various infections during the progressive stages of infection, and then at the end I will address the issue of the levels of viremia necessary to transmit. So, again, current risk update. In terms of looking at the patterns of viremia, I've divided up my presentation into sort of three categories of viruses. The first, the major transfusion-transmitted virus is for which routine screening is currently in place, and not only serologic screening but the advances with implementation of nucleic acid testing. Second is pathogenic viruses that we do not routinely screen for, and this includes the Herpes viruses as well as consideration of Parvo B19 in hepatitis A. And then a third category, which is viruses that actually are fairly prevalent and have been discovered over the last few years, but that have yet to have an established pathological relationship. But I think that as we'll look at these viruses, I'll suggest that these may be a good model for the next emerging agent, and they may be a tool to assess the efficacy of pathogen reduction methods in routine clinical practice because these viruses are being transmitted daily and, as you'll see, can exist in fairly high titers in donor units. Finally, at the end I'll briefly address the issue of the relationship of viremia levels to infectivity. This slide just summarizes the enormous progress over the last several decades through implementation of both enhanced donor selection criteria, but most important, really each of these dots represents advances attributable to improved screening. And the strategy of adding enhanced tests, as we've discovered viruses and built better tests, has clearly been successful, but I think one of the promises of pathogen reduction is that it might avoid the need to add a new test for each specific agent and allow us to feel more comfortable not adding new tests as new pathogens are discovered or as issues of residual transmission are documented. But I think as we'll come back to this at the very end and in the discussion, in my opinion I do not expect that pathogen reduction methods will allow us to discontinue any of these assays that are currently in place. In terms of the current risk, the classic way that we've been estimating risk over the last 10, 15 years has been using the so-called incidence window period model, where we quantify the rate of new infections in the donor pool. In the repeat donors, we assess the rate of seroconversions. We then estimate incidence overall and by factoring an increased rate of seroconversion for first-time donors. And then by knowing the durations of the window periods, either the serologic window period or with the introduction of NAT, the window periods that precede detection of RNA by either mini-pool or single-donation NAT, we can calculate out these risks of donations being given during the early window period. And as you can see here, we've, again, made enormous progress as we've moved from serologic testing to the introduction of mini-pool NAT, and could make slightly better--the blood slightly safer by moving to individual-donation NAT. But, really, we're at sort of the asymptotic level of these risk curves, with risks now in the range--with mini-pool NAT, in the range of 1 in 2 million for HIV and HCV. At present, we're not performing mini-pool NAT for HBV, in great part because if the very small incremental reduction that could be achieved over surface antigen test by mini-pool NAT. We could increase safety moderately for HBV and really minimally for HIV and HCV by moving to individual-donation NAT. And we'll talk through the window period data that really explains this in a few minutes. In addition, sort of a newer approach beyond the classic incidence window period model for estimating risk actually involves using the yield of mini-pool NAT as a tool to estimate risks associated with mini-pool NAT-negative units. And this is simply a strategy that uses the rate of detecting mini-pool NAT donations on a total-donation rate. This is actually Red Cross data, so we're picking up mini-pool NAT-positive units at 1 in 270,000 for HCV and about 1 in 5 million for HIV. And by knowing the relative durations of the window periods, the mini-pool NAT-positive pre-seroconversion window period, versus the total pre-mini-pool NAT window period or the individual-donation NAT window period, we can, in essence, adjust the rate of mini-pool NAT yield to estimate the risk of blood or the yield of individual-donation NAT. And I'm not going to go through this in any detail, but suffice it to say that these numbers now based on mini-pool NAT yield are virtually identical to the rates from the incidence window period model. So very reassuring, I think, that our risk is in the range of 1 in 2 million for each of these viruses, and that the yield of individual-donation NAT will be extremely low, probably about 1 in 5 million, for each virus. Now, as Roslyn emphasized, what we've done is to drive the risk of the viruses down to ranges of 1 in a million, and what we're left with is other problems from transfusions, such as bacteria that are, I think, much more important today. We also have a number of non-infectious complications, and so this is a slide from Sunny Dzik that I think puts into broader context where our problems lie. And I think today in transfusion medicine really infectious risks, with the exception of bacteria, are a very small contributor to the overall transfusion problem. Moving on now to the issue of viremia, for each of the agents we're sort of working towards understanding the sequential stages of infection and the levels of viremia that exist. And as we look at the current viruses and some of the new emerging agents, we'll be talking about these kinds of stages--in eclipse phase, which is the period following an infectious exposure, but before one can detect virus by testing the blood, even with highly sensitive nucleic acid methods. Then during that period of eclipse, we often see intermittent blips of viremia, and this has been termed recently "pre-ramp-up viremia." I'll illustrate that. Then there's this exponential explosion of viremia in the blood, sometimes associated with symptoms, as in primary HIV, but often asymptomatic, as in hepatitis C. We term this "ramp-up viremia." For some viruses, with, for example, HCV, there may be a prolonged plateau phase during which high-titer viremia exists for months before seroconversion. Other viruses, like HIV and HBV, tend to have a peak of viremia associated with seroconversion and then a clearance of that high-titer viremia. Often around the time of seroconversion, we'll see dramatic fluctuating viral load, often to negative values, only to be followed by a persistent viremia. And we've termed this "peri-seroconversion viremia." And then with seroconversion, people tend to either become persistent carriers, typical of hepatitis C or HIV, and they'll establish a viral load set point in the context of a seropositive state, and understanding that viral load is obviously important for targeting pathogen reduction methods. Some people will resolve the viremia, fairly common in hepatitis B and about 20 percent of HCV-infected people. Now, other kinds of odd events have been documented, what are termed "immuno-silent carriers," people who may have long-term persistent viremia in the absence of seroconversion. And there's also been documented examples where people can become transiently viremic, clear the viremia, and never seroconvert--in essence, an abortive infection or a successfully resolved infection perhaps by cell-mediate immune responses but without antibody conversion. This slide summarizes the story for HIV, and then I'll show a few examples. This is kind of the classic pattern. In a subset of people, probably 10 or 15 percent of the panels that we've studied, we can detect a blip of viremia, transiently detected--I'll illustrate that--about the time of exposure or within days of probable exposure. Then we have this explosive ramp-up phase, and we can quantify, as I'll illustrate, when individual-donation or mini-pool NAT or T24 antigen would detect viremia during ramp-up. We understand the dynamics of that ramp-up viremia quantitatively. Peak viremia for HIV typically is in the million to 100 million range, so very high viral loads exist transiently during this acute dissemination phase. And then usually the viremia stabilizes down to set points in the range of 102 to 105 genome equivalent in the setting of a chronic asymptomatic HIV carrier. This is just one example that illustrates a few points of a seroconversion panel that we've studied. This is a panel identified by Alpha Therapeutics and Bioclinical Partners. This person was actually not found to be seropositive until out here at day 16, with day zero defined as the first quantifiable RNA load sample. But these samples existed in the freezer to allow testing back and really careful study. What you see here is about two to three weeks prior to the early ramp-up viremia, this phenomenon of intermittent or erratic detection of a blip viremic event. In this case, you know, one bleed had one out of ten replicate high-sensitivity PCRs positive, and then the next seven out of eight. And then we continue into the eclipse phase when there's no detectable viremia for two or three weeks, and then very early phase ramp-up, with, again, three out of eight reps positive. This is a very sensitive assay. And then you enter this exponential ramp-up phase, this person peaking at about 3 million before they stabilize down. So a fairly typical case of early viremia. This is just to illustrate how we can quantify the ramp-up phase. This is, I think, something like 45 plasma donor panels where we can quantify the exponential increase in viral load, and you see many of these people will have peak viremias in the range of 10 million or greater. We can use that model, that regression line from the ramp-up phase, to estimate the viral loads at which different assays can detect the viremia. And this is really to emphasize here that in the setting of now mini-pool NAT, we can fairly confidently say that units should not be getting through the system that are mini-pool NAT-negative that have greater than about a thousand genome equivalents or viral copies per mL. And, clearly, if we moved individual-donation NAT, that viral load would be reduced to less than a hundred. So in the setting of contemporary screening, the residual viremia that should be present and associated with transmission should consistently be less than a thousand or, potentially with individual donation, at a hundred copies per mL. So, in a sense, the challenge for residual clean-up by pathogen reduction. This just illustrates that with progressive seroconversion, the viremia in a population analysis, again, tends to stabilize with the average viral load in the range of 104 to 105 in people who are chronically seropositive carriers. Moving on to hepatitis C, similar early in that there is a period of eclipse during which we see this intermittent viremia, pre-ramp-up blip viremia, followed by the explosive ramp-up phase. Again, we can precisely estimate when different markers would detect that early ramp-up viremia. But very different from HIV, with HCV there's a prolonged plateau phase. It lasts about two months, and during this period people are asymptomatic. Most of this phase, the ALT levels are normal. Only in the latter stages does the immune response, cellular response kick in and you get liver damage. The viral load during this plateau phase is enormous. It's in the range of 105 to 108 genomes, I believe, infectious copies per mL. So, really, a high viral load that, were we not to be doing RNA screening, would be, I think an enormous challenge for pathogen reduction to reproducibly eliminate. With seroconversion, most people remain chronic carriers, about 80 percent of people, with highly variable viral loads ranging from 100 to 107 in people who are persistently seropositive. Again, most of these people asymptomatic. Historically, of course, most HCV transmissions were occurring from donors who were seropositive chronic carriers. Just a few specific data slides. This is one of these plasma panels for HCV. This donor was detected by mini-pool NAT screening at NGI, and out at these time points, so this is a month or more before eventual seroconversion. And, again, by testing the stored plasma units, we could quantify the viral load during the early phase of plateau, this person in 107 to 108 copies per mL. These two points were contributing to the ramp-up phase analysis I'll show in a moment. But this slide is mostly to illustrate this phenomenon of the pre-ramp-up viremia. In this somewhat extreme example, for about two months prior to ramp-up we detected erratic virus only in two or three of the four replicate TMA--this is full input gen probe TMA assays. So it's this kind of low-level viremia that is probably accounting for most of the rare residual transmissions. Even with mini-pool NAT and even with individual-donation NAT, I think we could only erratically detect these low-level viremic very early phenomenon. For HCV, again, similar ramp-up phase. This is from about 40 or so plasma donor panels that had values during that very brisk early viremia. Again, many of these people achieving viral loads up as high as 108 per mL. And from this kind of analysis, again, we can model the detection of mini-pool or individual-donation NAT. In this example, we have the brisk ramp-up, and then just to illustrate one example of the plateau phase lasting in this case 46 days between the initial peak viremia and antibody conversion. This is a summary from Sue Stramer from the Red Cross, just to show the time from detection of donors as mini-pool NAT-positive to seroconversion. The yellow is the period during which these people are viremic. These are whole-blood donors picked up by mini-pool NAT at the Red Cross. Red is when they seroconvert, and you can see that most of these people seroconvert in the range of 50 to 60 days after initial detection. But this slide also illustrates two examples at the bottom of people who were immuno-silent carriers, picked up by mini-pool NAT, who remained mini-pool NAT-positive and antibody-negative for well over a year and well over two to three years. So these are examples of chronic carriers who never seroconvert. Again, just to illustrate, this is data from Dave Thomas. This is injection drug users who acquired HCV while in follow-up. Some of these people resolved infection and completely RNA-negative over prolonged periods. But this is just to illustrate that after seroconversion, most people, 80 percent of people, stabilize at fairly high viral loads. The average here is about 106 copies. So most seropositive HCV-infected patients have really pretty remarkable chronic viremia. One example for hepatitis B, this person is, again, a plasma donor panel who became infected with HBV. This is showing DNA load and surface antigen levels prior to seroconversion and then how this is resolved down associated with anti-surface and anti-core. The emphasis here is that, again, viral loads during this early peak viremia can be up in the range of a million per mL or greater. So quite a challenge for inactivation. Data again from Sue just showing in a population level the viral load over time as people develop surface antigen positivity, mostly here to emphasize again that in the later stages of primary viremia, viral loads in the range of a million or greater per mL are not unusual. As we've enhanced the sensitivity of antigen assays, the residual units that would be missed by antigen tests have lower-level viremia, down in the range of 10,000. And as we go into the phase of NAT testing for HBV, we further reduce that. This is illustrated in a slide from a study that should be submitted shortly from a collaboration of FDA and REDS looking at different antigen assays and the viral loads, and then as we get to mini-pool or individual-donation NAT, the viral loads that are still allowed to be, you know, released should be reduced well below 10,000 with contemporary antigen assays. And if we do implement NAT, we should be able to reduce the viral loads in the residual infected units down below 100 to even 10. Changing gears a little bit, we've talked about the three major viruses, talking about viruses that we do not routinely screen for, and starting with CMV, with CMV the frequency of transmission by CMV seropositive units is in the range of less than 1 percent. We just completed a study in collaboration with John Roback that looked at 1,000 donor samples to assess the frequency of detected viremia. These were tested by several serologic tests and by two DNA assays that had been validated through a previous multi-laboratory study of performance of assays. Each of these CMV, DNA, PCR assays had sensitivities of about 10 genome equivalent per 100,000 PBMCs. So these are very sensitive. In this example, we're targeting CMV-infected cells. In this study, the seroprevalence was about 42, 44 percent. When we completed the study with coded panels and retest panels, only two out of 416 seropositive donors, or 0.2 percent, were found to be viremic, and none of the seronegative donors were viremic. Importantly, in these seropositive donors who were infected, the viral loads were very, very low, less than 100 genome equivalent per 250,000 PBMCs. So this is in the setting of a kind of cross-sectional seropositive study. In a separate study that we've been doing looking at seroconverting donors, and specifically looking for plasma viremia, we completed a study that's been submitted that looked at a series of both serial samples from seropositive donors but, most importantly, focusing on seroconverting donors for CMV. And in this study, the only setting where we could find plasma viremia was in seroconverters. About 1 percent of these paired bleeds from seroconverters had detectable viremia. And in collaboration with Harvey Alter, we looked at serial samples from infected transfusion recipients who developed CMV, and we could quantify this period of primary infection where we could detect CMV DNA in the plasma often for two or three weeks, peaking at fairly low viral loads, at about a thousand per mL, during primary infection. Now, we don't have a lot--this is plasma viremia, because there are not a lot of cell samples from seroconverters for CMV stored in any repositories. So I contacted John Roback, who recently has done studies in the murine CMV model looking at acute infection, infecting these animals and then monitoring the blood, and this is showing copies of CMV DNA in the plasma following acute infection, and very similar to what we see in humans, there's an early period of plasma viremia, but quite low viral loads, in the range of 100 to 1,000 copies per mL. In stark contrast, though, they were able to study the frequency of CMV-infected cells, and they're finding viral loads as high as 105 to 107 per mL during acute infection. And this is probably the case in humans based on some limited human data, that in primary infection there may be quite high viral loads of infected cells. Another virus that's gained some attention recently is Herpes 8, which is the Kaposi's sarcoma virus. However, the studies that we've done, and also CDC has recently completed some large studies in Africa, suggested this virus, although there are a fair number of donors who are seropositive, the donors are consistently DNA-negative and that there is not transmission of this virus by transfusion. So this is a study that the REDS group has just completed with Phil Pellet from CDC which involved evaluating I think seven different laboratory assays for the prevalence of CMV antibody, and these tests detected CMV antibody ranging from 0.5 percent to 5 percent of the donors were said to be seropositive, and a latent class model analysis supported an overall estimated seroprevalence rate of about 3 percent in the REDS donor pool. However, all of these samples were PCR-negative, and, again, studies from CDC recently have shown in a large study in Africa that there does not appear to be any transmission of HHV-8 by blood transfusions. Parvo B19 is an important agent of concern to blood banks these days, a lot of considerations about the need to add screening for this by NAT methods. This is a non-enveloped virus that seems to be relatively refractory to inactivation. This virus tends to--it infects erythroid progenitors and causes anemia. It can also cause a problem in pregnant women and newborn infants. This has epidemic nature. Now, in terms of the viremia pattern, DNA is detected typically quite rapidly following infection, within four to eight days, and transiently for about seven days. IgM antibody is usually associated with resolution of the viremia. However, there have been examples of low-level persistence of viremia even in the setting of seroconversion. Now, very importantly, the viremia in acute infection is enormous, ranging as high as 1014 per mL. This is an example that Sue Stramer shared with me that shows one case of a plasma donor who was detected as viremic, and, again, in this example the viremia peaked at 1012 per mL, and this person remained actually PCR-positive at very low levels for months following that initial peak viremia. So, really, an enormous challenge for inactivation methods in terms of the viral load. Hepatitis A, classically foodborne hepatitis, but there are rare transfusion cases. Again, a virus of focus these days in terms of possibility NAT screening. Classic dogma suggests that hepatitis A is cleared very rapidly as seroconversion occurs, typically to undetectable levels within four to six weeks. There's classically no chronic carrier state, and people who have converted are thought to be immune for the rest of their lives. However, recent work from CDC has suggested that viremia may be prolonged and may persist in some people, so they had 13 individuals who acquired HAV while under HBV vaccine studies, and these people had viremia that preceded ALT by up to over a month. So before one would have any evidence of hepatitis, they were viremic. And in some cases, it lasted about three months. And the viremia levels, again, can be quite high, 105 to 107 per mL during the peak, and then staying up in the 1,000 or more for periods of months following conversion. The last group of viruses I want to briefly discuss are really viruses that over the last, you know, five to ten years have been, you know, sort of the focus of a lot of attention and then sort of put on the back burner because of the absence of confirmed evidence of clinical significance. However, I think that these viruses do present us with a potential model of efficacy, to study the efficacy of pathogen reduction, so I want to briefly summarize their characteristics. The first of these is hepatitis G virus, or GBV, discovered back about early 1990s and, again, thought to be a potential cause of residual non-A/non-B hepatitis, but subsequently shown to not be definitively hepatotropic or associated with any known disease. But, importantly, the rate of viremia in the donor pool is fairly high. In the range of 1 to 2 percent of U.S. volunteer donors are viremic, and much higher levels, 17 percent, in commercial plasma donors, clearly transmitted by blood transfusion. This is a slide from Harvey Alter looking at some transfusion-acquired HGV infections, and the point here is, again, very high viral loads. This person had a peak viremia of over 107 per mL. In this example, the person resolved the viremia associated with seroconversion, but there are also a number of people who remain persistent carriers, often with viral loads well in excess of a million per mL. These people are giving blood every day, and this virus is transmitted at close to 100 percent frequency. The TTV family of viruses, these are also non-enveloped viruses, so a problem in terms of potential other methods for inactivation. These are typically transmitted both by parenteral and fecal-oral routes. They do seem to be present in the liver but, again, there's not definitive proof that these cause any liver pathology or other diseases. Over the past five years or so, it's been shown that this is actually a very diverse sort of family of related viruses that include the more recently discovered SENV variant. In terms of prevalence in the donor pool--this is data from Harvey Alter--7.5 percent of the blood donations in the NIH, I guess, over the last, you know, ten or so years were positive for TTV DNA. In other studies from Japan, 50, 60 percent of donors are viremic for TTV. So a very prevalent virus. The viral loads in acute infection tend to be a little lower than with HGV, around 104 per mL, peak viremia, and although this slide shows some associations between that peak viremia and ALT elevation, larger studies have failed to prove that there's a causal relationship. Finally, SENV, again, it's a related virus to the TTV family. However, this is a virus that there is some suggestion may have hepatitis associations. In transfusion recipients we see, you know, 30 percent; in drug users, 60 percent. All of these viruses I've just spoken about, readily transmitted blood transfusion. In the U.S. donor setting, the prevalence of viremia is around 3 percent, but in other populations like in Japan, up to 30 percent of blood donors are viremic which SENV. This virus, again, does have an association with hepatitis. In Harvey Alter's group, 11 of 12 residual non-A/non-B hepatitis cases were positive for SENV viremia. However, Harvey's current conclusion is that this is not definitive evidence that there's a causal relationship. There's no question that this virus is readily transmitted by transfusion. And, again, one example of a recipient from Harvey's studies that had an early acute viremia, up over 1 million per mL, associated with some ALT elevations, and people can become persistent carriers with this virus as well. I want to just briefly touch on infectivity versus viremia to make the critical points that the level of virus detected in most of these studies by nucleic acid methods is probably a good reflection of the infectious titer of virus that needs to be eliminated by a pathogen reduction method. And the best data for this is really coming from our recent experience with breakthrough infections by NAT screening. For example, in this case reported in JAMA a couple years ago, a donor seroconverted to HIV in Singapore, and prior donations on lookback were demonstrated to transmit the virus. And when we went back and compared the ability to detect that infectious donation, which was in storage, using the newer NAT methodologies, what we found is that these methods were able to detect the infectious unit with full input assays. But as one diluted the samples out, mirroring the mini-pool screening context, these donations began to score negative. So this suggested--this was prior to widespread implementation of mini-pool NAT. This suggested that mini-pool NAT would miss some infectious units and that infectivity was present when very low viral loads existed. This is data from the San Antonio transmission case this last year where we similarly had plasma available from the implicated transmitting donation, and we were able to do the two licensed--the two commercialized NAT methods and show again that undiluted these methods could detect them. But as we diluted the sample out to the mini-pool sizes, erratic detection of the infection. So, again, clearly showing that HIV is transmissible when very, very low viral load is present. This has also been shown with hepatitis C. This was a case reported in Lancet a few years ago from Germany where a donation given eight weeks prior to C seroconversion was associated with transmission of HCV by the platelet but not the red cell concentrate, sort of illustrating that the volume of plasma in the unit is another variable that determines the total input of virus and, therefore, the infectivity. In this study, they were unable to detect the viremia present in that clearly implicated donation, and their conclusion was that they had transmission from a NAT-negative unit, emphasizing the point that even single-unit NAT may not prevent all transmissions. We actually acquired that sample and tested it and were able to show that we could detect very low level viremia, but only erratically. So this is very similar to the kind of blip viremia that we talked about earlier. So these studies, I think, show that window phase, antibody-negative units can transmit even when very low virus is present. Now, what about people who seroconverted? A lot of people have argued that once people seroconvert, there's neutralizing antibody, there's defective virus, so transmission should be, you know, suppressed by these phenomenon. But, actually, if you look at the data on real human transfusion transmission relative to viral load, there's strong evidence that very low level viremia transmits even in the setting of seropositive individuals. This was a study we did from the TSS cohort where we looked at transmission from seropositive donations and correlated the virus load in the unit and the duration of 4-degree storage of the unit with transmission. You probably can't see it very well, but there's really only a small percentage--over 90 percent of HIV seropositive units transmit. The only units that did not transmit are out here where they were stored for prolonged periods. If the sample wasn't in the fridge for more than a couple weeks, there was essentially a 100-percent transmission, irrespective of viral loads. And we're seeing transmissions even with units that had less than 100 genomes per mL in the setting of the seropositive person. And we have similar data for hepatitis C from the TTVS where we're seeing transmissions from seropositive donations, even if the viremia can only be detected by full input TMA analysis, 100-percent transmission, even with extremely very low viral load. So, essentially, if viremia is detectable, transmission is occurring at close to 100-percent rate. So, finally, just a summary slide to point out that, as we focus on these agents and we try to understand the viremia that is the challenge for pathogen reduction, for the major viruses where we're doing mini-pool NAT, the viral loads should be quite low, the challenge is low. But were we to not be doing mini-pool NAT, considering dropping NAT, were we to go to pathogen reduction, the viral challenge would be quite high. The window periods have viruses in concentrations of 106 to 108 per milliliter. In chronic infection, there's still very high viremia, so consideration of dropping antibody testing or NAT testing I think would result in a very high challenge viremia for the pathogen reduction methods. For the cell-associated virus, again, window period viremia, infected cells can be quite high, typically very low level viremia in the setting of antibody conversion; and, of course, leukoreduction for these viruses should reduce the viremia to low levels, although there's clear evidence that residual infectivity exists even in leukoreduced products. For these two viruses, hep A/B19, the viremia is enormous, 108 to 1012 during the window period, and there is increasing evidence that some people can become chronic carriers after seroconversion with low-level viremia. And then, finally, these sort of sentinel agents that I think serve as opportunity to further evaluate the efficacy of pathogen reduction. They're being transfused routinely today. These established chronic infections with viremia in the 104 to 106 range. So I think these viruses are an opportunity to really evaluate in clinical context the effectiveness of pathogen reduction. Thank you. [Applause.] DR. VOSTAL: Thank you, Dr. Busch. In the interest of time, we'll move on to the next talk, and you will have a chance to ask Dr. Busch when he heads up the viral inactivation discussion panel later on this morning. The next speaker is Dr. Leiby, and Dr. Leiby's a senior investigator at the Holland Lab at the American Red Cross, and he will cover the parasites and other pathogens that can be found in blood and transmitted by blood transfusion. DR. LEIBY: I'd like to thank the organizers for having me come here and talk to you about something slightly different, and that is, parasites and other pathogens that are also transmissible by blood transfusion. Now, we've already heard from Ros and Mike about the bacteria and the viruses, and we're going to venture into an area that's a little more gray, an area in which there's not as much information; and as far as pathogen inactivation and reduction, there's been relatively few studies, and those are the parasites and other pathogens. What I would like to do is give you an overview of those agents which we know are transmitted by blood transfusion, the relative levels of transmission cases which we do see, and then some of the issues that, as one begins to develop techniques for pathogen reduction, we have to take into consideration with these agents because they are in many ways different than what we see with viruses and bacteria. Broadly, the agents I'm going to talk about are the protozoa, the Rickettsia--which, in fact, are bacteria, but I think they fall better into a parasite classification--and, lastly, an agent of something that we might see in the future, I have included TSEs, now because we know they're transmitted by blood transfusion but because pathogen inactivation as a whole has a promise of being able to address agents for which we don't know about yet or which we might see in the future. I'd like to propose as we get into these agents, these parasites and others, that we, in fact, are raising the bar. In this sense, it's becoming more difficult as we address these agents to actually reduce them in blood and blood products. First of all, they are phylogenetically much more complex. They are carryouts as opposed to the agents we've been talking about so far this morning. In essence, these cells are more complex. Their outer cell surfaces in some cases are more complex, so they may in some instances be more difficult to kill or reduce. They also in many cases have very privileged intracellular niches. This is not to say that we don't find bacteria and viruses intracellularly, but some of these niches are rather unique, inside phagosomes, inside phagolysosomes. And so these may present problems which we don't see with the other agents. Many of these agents can fall under the classification of being emergent, so we know very little about them. We don't have tests available to tell us how often they are transmitted. We don't have models, animal models or other culture models, that allows us to measure how well the inactivation techniques are working. So as a group, then, these become a little bit more difficult to work with. Lastly, and perhaps most importantly, is the consideration of experimental models, because as we measure inactivation, not only do we need the agents to put into the blood products to test the inactivation strategies or reduction strategies, but we also have to have ways of measuring how well these strategies actually work. In the case of these, we don't have the simple culture methods we have for bacteria or viruses. We may have to go to actual animal models. We may have to go to unique culture systems in order to get the things that we need. In fact, as a whole, we need specific life-cycle stages, and as these organisms' parasitic life cycles sometimes are very complex, we need specific stages that may only be obtained from one source. As I said, the reduction and viability measurements are not so straightforward. We can't simply play it out on a culture plate to see if the bacteria grow, and we have to find a system that allows us to tell not only are the organisms there but, more importantly, are they still viable. And, lastly, we get into cellular survival and integrity, and this is a question that I'm sure will come up later with any of these techniques, what kind of damage do we see with the cells that are left after the reduction techniques. The agents I'd like to focus on today are the ones that I believe are the most important. The first one I'll talk about is Trypanosoma cruzi, which is the etiologic agent of Chagas' disease. Next I'll talk about the broader group of Plasmodium, which causes malaria. We'll talk a little bit about Babesia microti, which is closely related to Plasmodium, which causes babesiosis. Then we'll surely venture into the Rickettsia, which includes Ehrlichia and the recently renamed Anaplasma, which causes ehrlichiosis. Then lastly, as I said, not an agent that is known to transmit by blood transfusion, but we'll talk about new variant CJD. The first one I'll talk about is Trypanosoma cruzi, and I'll do the same thing virtually for all these agents. I'll provide a little bit of background, then some information on the seroprevalence or transfusion transmission and some of the problems which we may be facing with pathogen reduction. This is actually the agent of Chagas' disease, and as you can see here, it's a flagellated protozoan which is found extracellular. Unlike Ros and Mike, I can show you nice pretty pictures of these agents. It causes, most importantly, a chronic, asymptomatic, and untreatable infection, and it's actually those first two things that make it most important as an agent transmitted by blood transfusion. The donors are, in fact, asymptomatic, have no knowledge of the infection. They're also chronically infected. They're probably infected as children when they lived in an endemic country of Latin America. And for the rest of their lives, they are then capable of transmitting the infection during blood donation. As I said, it's endemic to portions of Mexico, Central America, and South America. Now, there's four primary routes of transmission. First of all is vectorial, which is actually the insect or the natural way in which it's transmitted. It's transmitted when the feces of the bug, which contains the infective stage, is rubbed into a bite wound or some other mucosal surface. It can also be transmitted congenitally, which is from mother to the unborn infant. Most recently, it's been reported--and this has been known in other places, but there is a case reported in MMWR in March of this year of transmission of Chagas' disease by organ transplant, in which a single organ donor transmitted the infection to three organ recipients, at least one of which died from Chagas' disease. Lastly, of course, the topic we're concerned about is blood transfusion. Now, as far as T. cruzi transmission by transfusion, if one looks at seroprevalence, in some at-risk populations the level is as high as 1 in 5,400 donors. In fact, Ira Shulman had a study in Los Angeles in which he put that level at about 1 in 550 donors. So it depends on which population you want to look at. Nationwide, it's probably about 1 in 25,000 donors in the U.S. are, in fact, infected with T. cruzi. And this number will vary depending on the number of at-risk donors in a given area, but I'd like to stress that there are no areas in this country where you would not find at-risk donors. The numbers just may vary. In a study we've done at the Red Cross, we found that of these seropositive donors, 63 percent of them are, in fact, parasitemic, so they actually have parasites in their blood, and we know that we are transfusing these parasites to donors. We've only seen those six transfusion cases as a whole in the U.S. and Canada. Certainly blood transfusion cases occur throughout Latin America, so this is nothing novel. The numbers seem low here. In these cases, they all involved immunocompromised recipients who got fulminant Chagas' disease. So what I always like to propose is that these six positive individuals in these six transfusion cases certainly served as sentinels, and they were the most obvious cases; whereas, most of them are either misdiagnosed, in many cases there have been underestimates of true transfusion transmission. Now, if one wants to look at pathogen reduction for this agent, what are the issues which one needs to consider? Well, first of all, unlike what Mike just talked about, the high levels of viremia--and Mike proposed that as an issue when you're trying to deal with inactivation. I'm going to propose just the opposite. With T. cruzi there is intermitted to low level infections. Sometimes you'll be able to measure parasitemia in the blood. Sometimes there are no parasites present. And, in fact, when parasites are present, they're present in extremely low numbers, maybe less than ten in a unit of blood. So when you're designing a technique that's designed to eliminate such a low number of organisms, it gets to be difficult in your experimental models to not only measure that but also to reduce those parasites that may be there. One advantage with this agent as opposed to the rest of the ones I'll show you, it's actually extracellular, so it's pretty easy to get at. It's found both in platelets and red cells. They're both capable of transmitting the infection. But when one gets into model systems of T. cruzi, things get a little bit more complicated. First of all, it requires metacyclic trypomastigotes--those are the infective stage of trypomastigotes--in order to really mimic what goes on in the blood bag. Those type of trypomastigotes can only be obtained through animal models or actually through sophisticated culture techniques. And then in the back end, the measurements of inactivation require two things, either culture or animal models, and there we must distinguish between what, in fact, are parasites that are present--one could easily measure by PCR, but that does not give you any measure of true viability. So one must consider the viability or the infected--whether or not these organisms that you see are, in fact, infective. Shifting to a slightly different agent, we'll talk about Babesia microti, which is the agent of human babesiosis, at least in the United States. It's an intracellular pathogen of red cells, and you can see some of the parasites in these red cells. It's a tick-borne zoonosis, transmitted by Ixodid ticks, more commonly called the black-legged or deer ticks. These are the same ticks that transmit not only Lyme disease but also vialichia (?), which I'll show you in a minute. As far as babesiosis in this country, there is local and regional distributions, primarily in New England, the upper Midwest, and then there's some agents that are similar in nature, which I'll show you, from the West Coast. When we get into Babesia and the rest of these agents, they all have what one describes as flu- or malaria-like symptoms, so they're very nondescript and difficult to identify individuals who are infected. Infections generally as asymptomatic and self-limiting, so unlike Chagas' disease, which is chronic and untreatable, in fact, babesiosis is quite readily treatable with antibiotics. The problem, though, is that in people who are either elderly, immunocompromised, or asplenic, the disease can be rather severe or even, in fact, fatal. Now, when one talks about transfusion transmission of this agent, the seroprevalence in studies that are published to date are somewhere about between 1 percent to 4.3 percent in donors. So it's not uncommon. Most seropositive donors are, in fact, also parasitemic, so we can measure the parasite in the red blood cells. And this parasitemia varies anywhere from a rather short period of a couple of months to greater than one year. And this suggests the possibility that, in fact, we may see chronic carriers or donors who have this infection for long periods of time and appear asymptomatic. Now, unlike Chagas' disease, there's been quite a few transfusion cases, 40 to 50 transfusion cases in the U.S. within the last 10 years. And we hear reports of transfusion cases almost monthly, so this is one that is becoming more and more prominent. In fact, this would certainly be an underestimate because most transfusion cases, again, are not recognized, particularly when you consider you have flu- and malaria-like symptoms. And they do cause some fatalities, so this is not a benign illness. Well, what about the issues with Babesia? How are they different? Well, first of all, it's an intracellular agent, so we have to get the--whatever reduction method we use has to be able to get inside the cells. The other one is that there are similar emergent agents, so-called WA-1, CA-1, and MO-1. These are Babesia-like agents which are found primarily on the West Coast, transmitted by a different tick. There's been at least two transfusion-transmitted cases of WA-1, but genetically these agents are somewhat different than Babesia microti. So the methodology used to reduce these agents may be slightly different and will have to be considered. These agents can be found, as you know, obviously in red cells, but due to red cell contamination, platelets have also been implicated in transfusion studies that have transmitted the agents. Now, when one speaks of model systems, we need, first of all, required--it requires infected human red cells. One could use animal models, in which we used hamster infected cells, but in a model system those would not be the priority, I think. The measurements of inactivation get a little more complicated because there, in fact, are no culture systems available for this parasite. And it's long been thought kind of Holy Grail. People have looked for culture systems, and at least for the human Babesia, these do not exist. So the bottom line, the measurement for inactivation would have to be in some type of animal model to see if these organisms are, in fact, really reduced or inactivated. Now, the close cousin of Babesia microti are the Plasmodium, the agents that cause malaria. These are also intracellular pathogens of red cells but also liver cells. But since we don't transfuse liver cells, we won't worry about those. It's mosquito-borne, primarily by Anopheline mosquitoes, and primarily limited to the tropics, so it's not something we face actively here in the United States. It causes flu-like symptoms, which often have a periodicity, meaning that the symptoms will come and change every two or three days or four days, depending on which agent it is. So it varies by infecting species. These are the two problems which we encounter in transfusion transmission in malaria. Although our current strategies of using questions about travel and where people have come from work quite well, it's when we get into the asymptomatic carriers or more likely the people who are semi-immune and actually have low levels of infection that we actually see transmission cases. The other complexity with the malaria is one designs techniques or pathogen reduction, and this is actually the list of the four human malarias: P. falciparum, P. vivax, P. malariae, and P. ovale. It's the rather complex life-cycle stages they have which might require different techniques or different constituents to remove them or reduce them, anywhere from these small ring stages to tropozooites to gametocytes. And all these organisms go through these similar complex life cycles. And as you can see, they're all intracellular as well. Well, what about transfusion transmission? In the U.S., seroprevalence is unknown because it really is not actively transmitted, so it would be very difficult to get any measure of what the actual seroprevalence is. Generally, in the U.S., there's about one or two transfusion cases per year, very low levels, and primarily, as I said before, these involves asymptomatic or semi-immune donors. Now, as far as issues that surround the development of tests for Plasmodium, these agents, again, are intracellular in red blood cells, as shown here. There's a variety of species and stages which would have to be considered. The model systems, again, would require infected human red cells, and then, again, the methods of inactivation, the measurements, would have to apply probably in culture systems or perhaps even animal models. Now, a group of agents perhaps you know less about than the others ones I've just mentioned are the Ehrlichia or the Anaplasma. These are the agents of human ehrlichiosis causing human monocytic ehrlichiosis and human granulocytic ehrlichiosis. There's actually even a newer form, Ehrlichia areini (?), that also is found among human leukocytes. By and large, these are all newly emergent agents, appeared in the 1990s. You can see some pictures of them here inside, some granulocytes, and actually the arrows are pointing at the parasites themselves, the small--actually, which they are--they form what are things called morulae, which is a Latin term referring to the raspberry/grape-like structure. Again, these are tick-borne Rickettsia transmitted by the same one that transmits Babesia and Lyme disease. It's intracellular, as I said, in leukocytes. And once again, we have the same nondescript flu-, malaria-like symptoms. And among patients who are infected, 10 to 20 percent have more serious complications. In fact, certain numbers of them also die as well. When one talks about transfusion transmission and when one talks about seroprevalence, there are very limited studies, virtually none published, but we think somewhere--in the recent one we have in press, the levels are anywhere from about 0.5 to 3.5 percent of blood donors in Connecticut and Wisconsin. There is one presumptive case of HGE transmission reported at AABB a few years ago, and in this case, the red cell units were implicated, thereby once again suggesting that it contained leukocytes that were infected. The problem with this agent, primarily because it's so new and emergent, is that it's underrecognized and also misdiagnosed. Well, what about the issues, again, as far as developing techniques for reduction? Well, first of all, the first issue, again, is that it's intracellular, but in this case, there may be some advantage, in fact, that universal leukoreduction may, in fact, remove most of these infected cells. As I said, it's newly emergent. In this case, it virtually lacks any test. We have no way of knowing how many donors are infected, how many are positive, and so forth. So there's a lot of issues that need to be worked on in order to get a better handle about how big a threat this agent actually is. There are several model systems. There are very good systems that allow ones to infect human leukocytes in culture, which works quite well, and that, as I get into measurements of inactivation, allows one to do it in cell culture. The point was that you need infected human leukocytes in order to do the experiments. The last one I'll include--and this falls into, I guess, the category of the other as far as my talk of parasites and others, and I throw this out not as an agent. I said before that actually we know to be transmitted by transfusion, but one which needs to be thought about, as well as other agents that are newly emerged, and also the promise itself of pathogen reduction to address those agents that we don't know about yet or we will see in the future, and this includes, as I said, transmissible spongiform encephalopathies, which we know includes new variant Creutzfeldt-Jakob disease. It's a prion protein, so I can't show you any neat pictures. It's an abnormal isoform of a cellular glycoprotein, thought to be transmitted at least initially through contaminated beef. That's how it got into humans, or at least we think. It invariably causes a fatal brain disease, the incubation period in years, and untreatable. There, again, anytime you get any of these agents that are incubated for years, we get into problems with our model systems for measuring inactivation or reduction. Well, a lot of these things are easy to answer this time. If one talks about new variant CJD and we talk about seroprevalence, well, we really don't know, and as far as blood transfusion, we have no evidence that it's transmitted by transfusion. But the issues that one would have to address if one wanted to look at this agent are complex. As I said, it's newly emerged, and the need for model systems, while there's a lack of tests and, as I already mentioned, very long incubation periods, it makes it difficult to look at these agents. So just to summarize parasites and other pathogens in this topic, they are indeed more complex organisms, and they require more complex model systems. And the fact that they're newly emerged makes the ability to find these systems or develop these systems slightly more difficult. Well, what about pathogen reduction itself and also the toxicity one may encounter? First of all, when you have an agent like some of these in which they have very low levels infection, it may be more difficult to eliminate them because we may have to increase the amount of whatever reduction technique or inactivation agent we use in order to actually see that they've been removed. So I propose that perhaps in some of these agents that only have one or two per milliliter--or a hundred milliliters, rather, it may be more difficult to measure their inactivation as opposed to the viruses which might refer to that have hundreds of thousands or millions. What about the cell viability? As we get rid of these agents which are found intracellularly, we may, in fact, also harm the cells. So those, I'm sure, are topics we'll hear more about today. And, lastly, what about diminishing returns? I raise this issue that in some of these more rare event cases in which we see maybe only a couple of cases of malaria per year or several cases of babesia per year, perhaps pathogen inactivation may be an overkill as a technique to limiting these agents where some of the more tried and true systems we already have may be the easier route and, in fact, the more cost-effective route. So I suggested that we may be treating many just to reduce a few cases. Thank you. [Applause.] DR. VOSTAL: Thank you, Dr. Leiby. Miraculously, we've managed to get back on time. Are there any questions for Dr. Leiby? [No response.] DR. VOSTAL: If not, we'll move on to the last talk in this session, and this will be presented by Dr. Richard Diamond. Dr. Diamond has just recently joined the Center for Biologics, and he's now the Assistant to the Deputy Director for Medicine, Dr. Jesse Goodman. Fortunately for us, Dr. Diamond has a 30-year history in academic medicine as an infectious disease specialist, most recently coming from the University of Boston School of Medicine. Dr. Diamond? DR. DIAMOND: It's a pleasure to have the opportunity to speak to you. I was asked to focus on bacterial infections, perhaps because there's a broad range of topics to address, and there's a need to keep things simple in terms of time. And there are critical areas to address in relation to bacterial infections that haven't been yet. And what I'd like to focus on in particular are what happens and what's the significance of organisms that are transfused and what about bacterial products; and as the prototype bacterial product, we'll try to address endotoxin in particular as the most clearly studied bacterial product which is known to have significant consequences in the human circulation. You've already heard that there have been a variety of studies of what happens when you add bacteria to various blood components. With red cells that are stored at 0 degrees--and a variety of different groups have added anywhere from one per mL to 100 per mL bacteria and seen how long it takes for inocula to grow, and really minimal inocula in that range of Yersinia enterocolitica, Enterobacter agglomerans, or Pseudomonas species, reach log phase growth and high levels of endotoxin, lipopolysaccharide by two to three weeks, with a generation time, even though it's slowed down, of anywhere from 15 to 22 hours. Other organisms that we would consider less pathogenic can also grow at these temperatures, even though we don't usually call them--consider them to be major risks for growing in solutions at low temperatures, but they certainly can. And, presumably, within this time period patients would receive significant inocula if they were transfused. Platelets, of course, being stored at room temperature, minimal inocula very quickly give you log phase growth, in less than one day in most of the studies that have been looked at, with stationary phase growth with numbers above five times 108 per mL, often about 1010 per mL in units that are four to five days--after four to five days of storage. Less pathogenic bacteria that we don't associate with human infections as much but certainly have caused them in recipients of contaminated blood, like Propionibacterium acnes, which is one of those organisms that can enter as a skin contaminant, may take a little bit longer to grow and have a slower doubling time when one artificially inoculates them into platelet samples, but certainly can reach very high numbers by four to five days and have caused significant infections. So how often does contaminated blood cause bacteremia? Well, if you look at the data--and this is from a really broad series of collected numbers in the literature--it's hard to come out with defined conclusion, but you can make some sense out of this. And I think the important message is that if one looks at the estimates of clinical bacteremia that occur from either red blood cell contamination or platelet contamination, it is only a small fraction of the--the rates of clinical bacteremia are really only a small fraction of the numbers of contaminated units that are transfused. In other words, even though these are high-risk patients, particularly those who are receiving platelets, they don't get infected most of the time when they receive contaminated blood. Some of them are receiving antibiotics, but not all of them are in that category at the time of transfusion. Even in high-risk patients, infusion of contaminated blood then usually has apparent effects. It's estimated, for example, after--the infection rate after receiving contaminated peripheral blood stem cell infusions is about 13.7 percent. Now, most of those patients are receiving antibiotics. Those numbers are colored, obviously, but the bottom-line message is that infusion of contaminated blood doesn't necessarily cause clinically apparent infection. If one looks at various rates of transmission of transfusion-related bacteremia and fatalities, it's hard to look at the multiple different series individually. This is from one series from CDC which I selected, not because it's necessarily better than any of the other ones in the literature, but because it's from the most recent time period, and one can at least say that these patients presumably receive state-of-the-art treatment or closer to it than some of the older studies. And one can see that the rates in that series of transmission of single-donor platelets, pool platelets, and red blood cell units, and the fatality rate in terms of percentages. So, again, it depends on the series, and it depends on who is and how the counting is done, but certainly not all contaminated transfusions lead to fatalities. So what variables affect clinical outcome? Obviously, the infecting organisms makes a difference. There's an enormous difference in strain and species virulence. However, common skin contaminants such as Bacillus species, Propionibacterium acnes, and so on can cause significant infections, and not only in those who we would think would be ultra-susceptible to infection, in other words, pancytopenia patients. The concentration and the rate of the bacterial infusion obviously makes a difference, and a variety of host factors, not all of which are very well defined, clearly neutropenia and immune status, but also a variety of other factors that have to do with cellular immune status, pre-existing antibodies to organisms or organism products in question, and receptors. It's known from mouse studies that there is significant genetic variation in receptors for Gram-negative organisms. The endotoxin receptor, the CD14 family of receptors, in terms of responses, is linked to toll-like receptors which seem to be responsible for transducing the signals to endotoxin exposure. And as this is being studies, it's very clear that at least different mouse strains have definable genetic variations in these receptors which can definitely determine outcome of infection, related receptors in the toll 2 family have been shown to be linked with survivability of staphylococcal sepsis. It's certainly highly likely that analogous genetic variation in humans might well be linked to some of the explanations for why seemingly normal people have very different outcomes to bacteremias, something that we who struggled with patients in clinical settings have been pondering over for a very long time. Now, also there's the issue of general clinical stability of patients, and that's where it makes it very difficult to interpret the very different series and different listings of fatality rates from the small series of recipients of contaminated blood that have been published. These can have major factors in both responses to a load of bacteria as well as to a load of bacterial products like endotoxin because the pre-existing status of the patients, the ongoing interactive factors are of major significance, and then therapy obviously makes a difference. During the transfusion, antibiotics can have a multiplicity of effects. First of all, antibiotics can raise the level of endotoxin liberated from bacteria 3- to 30-fold, depending on which series that one looks at. And this can be demonstrated in vitro and in vivo model systems. And, in addition, certain antibiotics bind endotoxin as well as binding and increasing clearance of organisms. Polymyxin has been used in vitro and in vivo to demonstrate for clearing endotoxin, and antibiotics that are in common use, some of the aminoglycosides like gentamycin, although they're not as potent binders of endotoxin as polymyxins, certainly are capable of binding endotoxin. And a variety of anti-inflammatories can have major effects on the cytokine responses that determine the effects and the outcome of bacterial and endotoxin exposure. And then how timely is recognition and specific treatment? From an analogous lesson of following immunocompromised patients for years, the old saw was, well, you better start treating these patients immediately when they get a fever and cover them with broad spectrum anti-Gram-negative coverage because these infections are rapidly fatal in neutropenic patients. And Gram-positive infections cause significant morbidity and some fatalities, but not as rapidly. If you take that as a lesson, the lesson here is there are major differences in host susceptibility, but also the lesson in recent years is if we look at that group of patients, they all are treated so promptly with broad spectrum Gram-negative coverage that it's hard to find a series where the survival of Gram-negative bacteremia is less than 90 percent in those patients, so it's hard to do meaningful comparisons of agents or regimens. So host factors do play a major role, and, in fact, if one looks at all these series, what comes out is, if you get contaminated red cells, your major risk of post-transfusion bacteremia correlates with pancytopenia and immunosuppressive therapies. Surprise, surprise, that doesn't tell you anything you didn't know or anything that's terribly helpful. There is a presumed relationship to impairment of bacterial clearance mechanisms which is responsible for this difference. And these patients are--one commonly sees that clinically significant infections can be initiated by less virulent bacteria and lower bacterial inocula. These have been shown in experimental models and to a lesser extent and more indirectly but with some compelling evidence in clinical and epidemiologic studies. In transplant recipients, for instance, symptomatic bacteria occurred in one out of 1,700 pooled random donor platelet unit infusions that were transfused, and that's one out of 350 transfusions in these patients. When one looks at it in that context, this is happening pretty often in susceptible groups of patients. But, really, is there a definable threshold infectious dose? The short answer is, well, you can't really tell in an individual patient. Infectious doses and attack rates are hard to determine in clinical and epidemiological studies. You don't exactly know what went in. There aren't good measures of inocula. We don't tend to measure them in advance. Most bacteremia-associated contaminated platelet infusions probably contain between 106 and 1010 bacterial per mL. Of course, if somebody has an immediate reaction, which they might very well to a massively contaminated unit, they don't get the whole unit. And most of the reports don't characterize exactly what they got, so it's very difficult to tell even if you knew what was in the bag at the time of the infusion, and there's precious little information that characterizes things that much. You don't know what the real dose was that was given to the patient, and, clearly, many patients get infected by units that are less contaminated than that. And, in fact, as you've already heard this morning and was intimated, it's much harder to tell when that happens, because the paradigm has been that contaminated blood and measurement of transfusion reactions are limited to immediate reactions. And we know from human studies and from animal models that inoculation intravenously with contaminated solutions of any kind, not just blood, the effects may be delayed and delayed significantly in animal models, the lower the inoculum one gives, and depending on the host factors that are present. So bacteremia after continued IV infusion we know can occur with 100 to 500 milliliter infusions with 105--between 105 and 107 bacteria per mL in maybe 10 to 30 percent of non-immunosuppressed patients based on series of not blood contaminated infusions but contaminated intravenous solutions where it's been possible to at least come up with a little bit of an idea of the range of contamination within the solution. But even then the information that you'd want to know in detail what the true inocula were and what the exposures were and what the rate of infusions were are not available clinically and epidemiologically for you to make hard and fast conclusions about that. But what does come out of this is that the mortality rates, considering the amount of bacteria infused, are not all that high. Now, they're higher than anybody would like to see, but considering the human experimentation of infusion of intravenous bacteria, these patients--most of these patients survive. Lower numbers of bacteria certainly can cause bacteremia in immunosuppressed patients, and probably--and, again, very difficult to judge and extrapolate from the epidemiological data that are available, but from some anecdotal series, 10 to 100 mL infusions intravenously probably containing anywhere from 102 to 105 bacteria per mL have initiated bacteremia, but not invariably. If one looks at then something that we would hope would be more definable, what are the reactions to endotoxin, lipopolysaccharide? This has been well defined in non-human primates, and there are some very nice models there. If one looks at septic shock, E. coli septic shock models in Rhesus monkeys, you can see the levels of infusion, 375 to 500 microgram per kilogram per hour for eight hours, or in baboons, one and a half milligrams per kilogram over ten minutes. The minimum lethal dose seems to be on the order of three to six times 106 endotoxin units intravenously, and that will kill about 15 to 30 percent. E. coli sepsis, non-fatal sepsis with 5 to 40 times 108 colony-forming units over two hours, four times 1010 is lethal to baboons uniformly. But humans are maybe 100 times more sensitive to lipopolysaccharide than any of these primate models, so they're of limited value in judging what will happen when we infuse endotoxin or bacteria. So what are the responses to endotoxin in humans? The minimum dose, in the range of 0.1 to 0.5 nanograms per kilogram, that will cause fever has been defined. That's 0.1 to 0.5 endotoxin units. If one infuses 2 to 4 nanograms per kilogram IV, which were the standard doses for normal volunteer studies in humans--and these are healthy, normal volunteers--one gets levels within 15 minutes to an hour that peak between 16 to 240 picograms per mL. These are all healthy volunteers, and they actually precede the peak effects. Signs and symptoms that develop are fever, chills, nausea, myalgia, headache, leukocytosis, and hypotension. They're somewhat delayed in onset after infusion, and that's presumably because they are related to the physiological responses that have to do first with primary release of TNF alpha and then with a whole slew of other cytokines and mediators, and there's also activation of the clotting and fibrinolysis system, although in these normal volunteers, this does not proceed to clinically evident BIC. So there are ways of controlling these processes. And then when one gets up to 20 nanograms per mL or 100 endotoxin units, you get into a range of serious toxicity. This is sort of a guess from--again, from epidemiologic data because this isn't the kind of study that is approved for general human studies use. So what about endotoxemia? First of all, it's not always detectable in septic shock. About 30 to 50 percent of septic shock patients where endotoxin levels are measured have detectable endotoxin, and usually it's on the order of 400 picograms per mL or above if present. But continuous exposure differs from a bolus dose, and the bolus doses that we see in volunteers aren't the same kind of exposure--it may be similar to what happens if you kill off all the bacteria in a blood unit, and then infuse it with the endotoxin that's left behind. But it's not going to be analogous to a situation of ongoing sepsis where the complex mechanisms involving endotoxin clearance balanced with ongoing effects on the immune and inflammatory responses are going to change the output of cytokines and the cytokine profile and inflammatory mediator profiles drastically, not to mention the complexities of inter-current factors in unstable patients where endotoxin may have very unpredictable interactions with other ongoing processes that affect the inflammatory and immune mechanisms. And high levels can occur without bacteremia, for instance, in ARDS patients, pancreatitis, cirrhosis, maintenance hemodialysis, and, interestingly enough, in one study ultra-marathon running. What's interesting there is they measured--some runners had over a thousand picograms per mL after running 81 kilometers. I'm not sure--I was dumb enough to run about ten marathons years ago, but I couldn't even conceive of doing that. But all they felt was some nausea and vomiting, which I think you have a right to feel after 81K, associated with this level of endotoxin. So depending on the circumstances, it's very hard to predict what effect a given level of endotoxin is going to have in an individual patient by itself, and that other factors presumably have major effects on modulating what responses you're going to see. It just makes it hard to predict what a given measurement or level means. So levels correlate very inconsistently with outcome, and complex dynamic factors determine what the levels are. So if you look at one case, for example, of an outbreak of 11 cases that were reviewed by CDC of Yersinia enterocolitica sepsis after contaminated red cell transfusions, the mortality was 45 percent. Endotoxin levels were tested in 5 out of the 11 patients. The median was 11,645 nanograms per mL. That's a few logs higher than the last slide we were talking about in the levels for what's commonly described in sepsis with a range of 3,500 to over 17,000 nanograms per mL and most were receiving antibiotics, which, as you recall, can increase 3 to 20 times or more the endotoxin levels in the circulation. What effect that has is unclear. And at least some patients, in this case the majority, 6 out of 11, can survive with enormous circulating endotoxin levels. That's about the only thing I can make out of this. So we've got examples of bacteremia--even though most cases of contaminated blood products that lead to bacteremia seem to be relatable to skin contamination, some have been documented as related to episodes of occult bacteremia in donors. And, in fact, I remember as a clinical associate making rounds in the clinical center in the early 1970s an outbreak on the childhood leukemia service of four cases of Salmonella choleraesuis bacteremia. The only common feature these people had, these kids had, was they had received platelet units from the same donor, and traced back to the donor, and the donor had an occult Staph--excuse me, Salmonella osteomyelitis, which presumably led to intermittent bacteremia in very low levels that one would have to go to some extremes in those ancient days 30 years ago to detect. So platelets are particularly vulnerable because of the storage temperatures, but also because so many organisms are adherent to platelets and can adhere very readily. In fact, I spent the last few years of my laboratory research career looking at some of the antimicrobial effects of platelets and some of the interaction of platelets with a variety of different organisms. It's striking how adherent platelets are and how readily platelets could actually concentrate organisms as one concentrated platelets. And many organisms can cause prolonged bacteremia without concomitant symptoms. We've heard about some of them. I might mention Bartonella, Borrelia, Brucella, we've heard about Ehrlichia, Treponema pallidum, Rickettsia. There's currently a low prevalence of these organisms in donors, we think, but our lessons about emerging infections in recent years are that things can change. And what that tells us, I think, is that we need to set up epidemiologic screening methods and be ready to change whatever paradigms we have for risk management when we are trying to eliminate pathogens from blood. It's not going to be possible ever to eliminate all risk of everything, and the relative risks for any given group of pathogens is likely to change over time, and we'll need epidemiologic data to be able to respond promptly to how this is taking place. Again, many organisms adhere to platelets and red cells, and some are invasive. These are red cells that have been invaded by Bartonella. You can see--actually, the slide doesn't come across that well, but there's some stainable material inside the red cells that represent Bartonella. This is a chronic infection that has been recently recognized as being more common than we had thought in the past and where people can have asymptomatic bacteremias that go on for months. We haven't had a case of transfusion-transmitted syphilis since 1969. Syphilis is very low prevalence in the population, and there are screening tests for people for this infection. On the other hand, there are other agents that are more common and may become more common yet, and we've heard of a number of examples this morning where there could be rising areas of concern and where there's important needs for epidemiologic data so that risks can be accurately estimated. Even with infections due to agents that invade and reside in leukocytes, depletion of leukocytes from blood doesn't preclude transmission by transfusion. This is human granulocytic Ehrlichiosis infecting a leukocyte. It's been shown with Ehrlichia and a number of other organisms that reside intracellularly in leukocytes that one can, even after depletion of leukocytes, demonstrate organisms in supernate, and that these organisms might well be transmitted by transfusions that were leukocyte-depleted. So that by itself might not be sufficient. So when one considers the pathogenic mechanisms that allow cellular adherence, that allow for cellular penetration, and the various localizations and the tremendous diversity of pathogens, it is an enormous challenge to begin to pick and choose how to decontaminate transfused blood components. It is also critical to keep in mind that endotoxin has some important lessons and that it is not the only bacterial product that has clinically consequential responses and provokes clinically consequential responses which may be potentially life-threatening in susceptible patients. What this means is that any mechanisms that are used for decontamination presumably will release bacterial products. It is essential to know what clinical effects these released bacterial products will have, and it is naive to think that mere elimination of detectable live agents, no matter how effective, will necessarily make transfusion of blood products safe. The consideration of the release of products of the organisms and the effects that they'll have clinically is a critical component that has to be addressed. So what can we say? Is there a safe level of bacterial challenge? It depends. Bacterial species and strain virulence obviously make a difference. The host immune status and clinical stability makes a difference. Host genetic factors, treatment. The short answer is for most patients probably, but we can't predict for sure which ones or how much any given individual will be able to clear without any adverse consequences. Is there a safe dose of endotoxin? In a defined bolus dose in volunteers, yes, there is. That's normal, healthy people who get a bolus of endotoxin. During ongoing sepsis, with ongoing release and unpredictable clearance of endotoxin, expect the unpredictable. Outcomes probably depend on the duration of endotoxin exposure and the nature of interaction with other factors because endotoxin can alter a host of physiological and pathological processes in cells and tissues as well as responses to drugs, mediators, cytokines, and other stimuli. And even safe doses for healthy volunteers may harm unstable immunocompromised patients. So since I haven't offered you anything helpful, I'd be glad to answer any questions. [Applause.] PARTICIPANT: I have a comment and a question. The comment is that the lipopolysaccharides from different Gram-negative organisms have different structures and may have different effects, and we and others have done studies to show that in human, if you use LPS from Brucella abortus and compare it to LPS from E. coli, for example, the LPS from Brucella abortus has a two- to three-log lower potency than the LPS from E. coli. So I think all Gram-negative organisms and all the LPSs are not equal. The question I have--well, I have a question to follow that. The question I have is related to your comment, which I think is very important, that Gram-positive organisms and Gram-negative organisms, when you inactivate and you may still have cell wall components that may be very active, and there's been a lot of recent literature about how even Gram-positive cell wall components can activate toll-like receptors and stimulate TNF release. DR. DIAMOND: Absolutely, and there's a limited amount of time to discuss all these things, but certainly there are--just as there are strain differences between bacteria of the same species, the variations in endotoxin structure actually that have relation to biological effects have been reasonably well characterized on a molecular basis. And certainly there are enormous differences in endotoxins between different species, and in peptidoglycans and other bacterial products between different species. The point is not about any specific one, but I think more in a generic sense that all of these compounds are going to be there. There are too many that might be of consequence, and what we desperately need are some sort of biological correlates that we can begin to use as an index for what we may be doing when we're making any changes in organisms, and then infusing what we have left. DR. VOSTAL: Thank you, Dr. Diamond, for tackling that difficult subject, and also thank you for pointing out another reason why not to sign up for those ultra-marathons. We're now up against a break, so if we could take a short break and come back at 10:45, we'll get started on the other sessions. [Recess.] DR. VOSTAL: Okay. Now we're going to move into our next session, and this session will cover the methods of decontamination or of pathogen reduction. Our first speaker will be Dr. Steve Wagner, who's the director of cell therapy and blood cell therapy development at the American Red Cross, Holland Labs. And Dr. Wagner has actually been helping us in planning and carrying out this workshop, so we really appreciate his input. So please welcome Dr. Wagner. DR. WAGNER: Thank you very much. I'm very grateful to be able to help plan this workshop and also to participate in it, and my task today is to give a general overview on methodologies for pathogen reduction. In addition, what I'd like to do is talk about a few points that I think might be important in trying to analyze some of the potential benefits and risks that are involved in these methods. We've heard from different speakers some of the rationales for inactivation. They are to eliminate or to reduce the residual infectivity in blood from pathogens. We all know what the danger of pooled products are where infections might be disseminated to many individuals, and certainly inactivation techniques might help in that regard. Inactivation techniques also might provide an additional layer of safety on top of testing and other things that we might do. There are some agents, as Dr. Busch and David Leiby have indicated, that we currently have no test for, and so pathogen inactivation might provide some additional safety in those situations. We haven't talked very much about--and I certainly won't talk much about the fact that some agents mutate and there are different varieties, and certainly this has been the case with HIV. And there have been some--there are differences in terms of the serological characteristics of different strains of HCV. And so variant agents which may not be amenable to current tests might be another category of agents that might be--where safety might be improved with inactivation techniques. And people have talked about new agents, and, of course, there's public and political expectations of a zero-risk blood supply. With all that said, pathogen inactivation is not necessarily a simple problem to solve. There are inherent challenges. Pathogens may appear in different compartments. Many of the pathogens have different structures. Dr. Leiby talked about that. And, in particular, in the virus families, there are some non-enveloped viruses whose capsid proteins are so tightly interdigitated that it's very hard for small molecules to permeate the capsid, and, in particular, the picornaviruses, of which hepatitis A is in that family, the capsid proteins are so tightly interdigitated that even a cesium molecule is unable to get inside the capsid. And so you might expect that these types of agents would be very difficult to inactivate. The processing that occurs with pathogen inactivation may reduce cellular yields. If you have to move blood from one container to another that might have different removal devices or other things, there are going to be some cells left behind. And so we shouldn't expect to have 100 percent of the cells that we started with after the inactivation technique. In addition, most of these agents work by targeting nucleic acids, but like most things in life, nothing is perfect, and there are unwanted side reactions from the treatments. And these side reactions may lead to los