Sepsis complicating platelet transfusions

Transcription

1 College of American Pathologists Laboratory Accreditation Checklist Item TRM Sepsis complicating platelet transfusions has been reported in many investigations, 1 10 including a case report and literature survey by Burns and Werch 11 that was recently published in the Archives of Pathology & Laboratory Medicine. In an effort to reduce the occurrence of septic transfusion reactions due to bacterially contaminated platelets, in December 2002 the College of American Pathologists (CAP) Laboratory Accreditation Program (LAP; a voluntary program of professional peer review and education that uses a series of checklists to assess the quality of clinical laboratory services and compliance with established performance standards) added a phase I item to the Laboratory Accreditation Checklist, TRM.44955, which asked the question, Does the laboratory have a system to detect the presence of bacteria in Platelet components? 12 However, the blood banking community did not hold a unanimous opinion that the CAP LAP should retain checklist item TRM Some colleagues have challenged the rationale behind item TRM and have posed questions and/or raised issues for discussion by the CAP Transfusion Medicine Resource Committee (TMRC), requesting that Accepted for publication May 14, From the Keck School of Medicine, University of Southern California, and Department of Pathology, Los Angeles County University of Southern California Medical Center, Los Angeles. Corresponding author: Ira A. Shulman, MD, Department of Pathology, Los Angeles County University of Southern California Medical Center, 1200 N State St, Room 2900, Los Angeles, CA ( ishulman@usc.edu). Reprints not available from the author. Phase I Requirement on Bacterial Detection in Platelets Ira A. Shulman, MD the TMRC reconsider whether TRM should remain in the CAP LAP Transfusion Medicine Checklist. The CAP TMRC has deliberated carefully over all submitted questions and comments pertaining to TRM and has reached a consensus that the item should continue to be included in the checklist. This article presents a sampling of questions and issues thathavebeenreviewedbythecap TMRC, as well as a summary of the committee s current consensus opinions. 1. The clinical experience of physicians does not reflect the incidence numbers reported for bacterial contamination; is this due to underreporting or misdiagnosis? It is the consensus of the TMRC that cases of bacterial contamination induced adverse reactions (including sepsis) are underreported owing to lack of recognition and/or misdiagnosis. 6 In the context of evidencebased medicine, clinical experience is a poor method for measuring most transfusion-related events accurately. If clinical experience were the primary method for measuring transfusion-related events, the true rate of viral infections or other transfusion reactions would not be known. Individual physicians do not encounter a sufficient number of patients to accurately quantify the risk. In addition, anecdotal experience and retrospective reports are unlikely to portray an accurate picture. Individuals who have not personally observed reactions to bacterially contaminated platelets perceive that the rate of clinical reactions is low. Alternatively, prospective studies of bacteria-induced reactions based on predefined criteria are most likely to provide accurate estimates on incidence. This is not a problem unique to bacterial contamination of blood components. In carefully monitored studies, such as those by the Centers for Disease Control and Prevention, Red Cross, American Association of Blood Banks, 3 at Hopkins, 1 and at the University Hospitals of Cleveland, 5 septic outcomes and deaths exceed those attributable to other transfusion-transmitted infections. Clinical experience from those institutions with a heightened awareness of the problem suggests that morbidity and mortality occur with a relatively high frequency. In other words, those institutions that look for the problem seem to find it. From the currently available data, it is impossible to accurately determine the true prevalence of transfusion-associated septic events. Most contaminated platelets are unlikely to cause clinically evident septic reactions. Moreover, even when transfusion-associated septic events are clinically evident, they often are not recognized as such. For example, a recent transfusion-associated septic case due to salmonella 13 would not have been reported as transfusion related had the blood donor not provided 2 separate contaminated platelet apheresis units, each causing a distinct transfusion-associated septic event in a different patient. The data reported by Ness et al 1 and from the French Hémovigilence system 8,9 suggest that the rate of transfusion-associated septic events per platelet unit transfused is 1: If one uses criteria for investigation in which febrile or other serious reactions to platelets trigger a culture, the- 958 Arch Pathol Lab Med Vol 128, September 2004 Shulman

2 transfusion-associated sepsis rate is 1: for single donor platelets or 1:2000 for a pool of 6 random donor platelets. Multiple reports have suggested that 1:1000 to 1:3000 platelet units are bacterially contaminated If it were assumed that 4 million platelet units are transfused annually in the United States (1 million apheresis platelets and 3 million whole blood derived platelet concentrates), it would be expected that as many as 4000 bacterially contaminated units would be transfused. Of these contaminated infusions, perhaps 1 in 6 to 1 in 4 would be expected to result in clinical sepsis (up to 1000 cases) and perhaps 1 in 5 to 1 in 3 would result in death (up to 333 deaths per year). 1,5 This extrapolates to a risk of death from a bacterially contaminated platelet unit of 1: to 1: Obviously, this high rate of death is not what is being reported, suggesting that either the estimates are incorrect or reporting is incomplete. The validity of these high estimates is consistent with clinical observations from university hospitals. Johns Hopkins reported a fatality rate of 1: with pooled whole blood derived platelets 1 and 1: with single donor apheresis platelets. University Hospitals of Cleveland 5 observed a fatality risk of approximately 1: per platelet unit. 2. The current stated risk of bacterial contamination of 1:1000 to 1: 3000 per unit may not be associated with adverse effect because patients are often on antibiotics, bacterial loads are often low, and patients can clear bacteria without clinical consequence. It is a valid point that the current stated risk of bacterial contamination of 1:1000 to 1:3000 per unit may not be associated with a similar degree of adverse risk, and that this measure is not an indicator of the actual frequency of serious outcomes. The rate of clinical reactions is certainly lower than the rate of transfusion of contaminated units. As stated, most contaminated platelet units do not cause a clinically evident septic event (if they did, the occurrence of septic complications should be much higher than what is reported), possibly because the bacterial load in many contaminated units is low and the organisms contaminating the platelets are of variable virulence, and/or the platelet recipient is already on antibiotics for associated clinical conditions. However, the issue of antibiotics is complex, as there are no controlled data showing that a clinical reaction following transfusion of bacterially contaminated platelets will be lessened by preexisting antibiotic therapy or that subsequent bacteremia will be lessened if the patient is already on an antibiotic regimen. In addition, while it is true that some bacteria are slow-growing and some may be of little clinical significance, an antibiotic regimen would not be expected to protect against a transfusion reaction that is induced by the infusion of endotoxin from a contaminating gram-negative organism. Many deaths attributed to transfusion-associated sepsis have been related to gram-negative organisms. 11 It is also possible that a patient s underlying medical condition may make it easy for clinicians and transfusion medicine specialists to overlook a connection between a contaminated unit and a transfusion-induced bacteremia. Such connections may be causal far more often than previously recognized. For example, in a prospective study of 3584 platelet transfusions given to 161 patients after bone marrow transplantation, 7 a septic workup was initiated for a temperature rise greater than 2 C above the pretransfusion value within 24 hours of platelet transfusion or a temperature rise of more than 1 C that was associated with chills and rigor. A diagnosis of bacteremia following platelet transfusion was made only when the pairs of isolates from both the blood and the platelet bags were identical with respect to their biochemical profile, antibiotic sensitivity, serotyping, or ribotyping. Thirty-seven febrile reactions, as defined above, occurred. Bacteremia subsequent to platelet transfusion was diagnosed in 10 patients in conjunction with 27% (95% confidence interval [CI], 15% 43%) of the febrile reactions. In a subgroup of 19 patients with a temperature rise of more than 2 C, the risk of bacteremia subsequent to platelet transfusion was 42% (95% CI, 23% 64%). Septic shock occurred in 4 of the 10 bacteremic patients. A rapid diagnosis was possible because the involved bacteria were demonstrated by direct Gram stain of the samples taken from the platelet bags of all 10 patients. Finally, despite lack of appreciated symptomatology, bacteremia in platelet transfusion recipients, many of whom are immunocompromised, should be considered a serious event The possibility of false-positive results due to contamination in the culture process may affect the numbers stated in item 2. When performing a culture on a unit of platelets, false positives may occur. While this is a valid consideration, multiple studies, such as those already discussed, have assessed the possibility of contamination during the culture process and have consistently found a contamination rate of approximately 1:1000 to 1: How often is culture positivity associated with recipient infection and/or reaction? How serious are the reactions? How often do they result in death? Johns Hopkins reported a fatality rate of 1: with pooled whole blood derived platelets 1 and 1: with single donor apheresis platelets. University Hospitals of Cleveland 5 observed a fatality risk of approximately 1: per platelet unit. One of the difficult variables to consider when determining the recipient infection and/or adverse reaction rate is the mix of single donor versus pooled random donor platelets. Data from Johns Hopkins suggest that a pool of platelet units made from 4 to 6 pooled individual units has an increased risk of contamination when compared with a unit of single donor apheresis platelets, 5 although data from the Bacterial Contamination Study (BaCon) do not support that conclusion. 3 Many reported transfusion-associated septic reactions are clinically severe, resulting in a need for placement in an intensive care unit and prolonged hospitalizations. Less severe reactions, such as the development of fever in a leukocytopenic patient with malignancy are not insignificant and may cause morbidity for the patient, given the clinical setting. 5. The risk of platelet shortages may outweigh any putative benefit of 960 Arch Pathol Lab Med Vol 128, September 2004 Shulman

3 testing for bacterial contamination; holding product while waiting for culture results may result in short date inventory management problems, particularly if the hospital is at some distance from the supplier. It is true that problems with platelet inventory management and even increased wastage may occur following a shift to a system designed to detect bacterial contamination of platelets. We will not know for certain what the impact of TRM will have on the availability of platelets until experience is gained. However, this is not the first time that there has been a required change in platelet inventory management that has had the potential to impact product availability. Previously, the Food and Drug Administration (FDA) mandated a reduced shelf life for platelets from 7 to 5 days (making them more likely to expire during storage and creating inventory management challenges) to reduce the risk of bacterial contamination of platelets and sepsis in recipients of the older stored platelets. Blood banks adjusted to the reduced storage period of platelets after the FDA mandated a 5-day maximum. However, in the present situation, new automated platelet collection systems have recently been approved that are satisfactory for a 7-day expiration of platelets in some areas of Europe, and such systems, if implemented in the United States, may improve platelet availability. In fact, some automated platelet collection systems have been cleared by the FDA for a 7-day expiration of platelets, provided the collected platelets undergo FDA-approved testing for bacterial contamination using a method cleared by the FDA for bacterial screening. Currently, the FDA has not cleared any method for bacterial screening of platelet products in the United States. Methods are only cleared for quality control testing of platelet products. The FDA has defined quality control (even if applied to 100% of the blood center s inventory) as a test that may be conducted on a platelet product, but the result of which cannot be used to extend platelet storage beyond 5 days. In contrast, when a test is cleared for bacterial screening, the test results may be used to extend the shelf life of platelet products beyond 5 days. Quality control methods and bacterial screening are, therefore, not equivalent per the definitions used by the FDA. 15 It is possible to use a scheme of platelet management that minimizes the likelihood that compliance with TRM will reduce platelet supplies. For example, the approach at Dartmouth-Hitchcock Medical Center is to perform culturing of platelets on a routine basis, per the directional insert of the equipment or method being used. However, as soon as the culture is taken, the unit is placed into available inventory, to minimize any delay in distributing the unit for transfusions. By the time the unit is required for transfusion, it is likely that a truly positive unit will have already been found to be positive in the culture system. Thus, testing does not necessarily have to disrupt the supply chain (James P. AuBuchon, MD, written communication, May 2, 2004). Indeed, United Blood Services has implemented Pall BDS testing of platelets at their Blood Centers of the Pacific center in San Francisco, Calif (in June 2003) and a BacT/ALERT system (in September 2003) at other United Blood Services centers with a minimal change in platelet outdating (J. Daniel Connor, BS, MM, CPA, written communication, April 5, 2004). When addressing inventory management issues, it is also important to recognize that the CAP checklist item TRM does not require bacterial culture of platelets, so that if a nonculture approach is taken, there should be no delay in providing platelets for transfusion. First of all, TRM asks, Does the laboratory have a system to detect the presence of bacteria in Platelet components? Thus, the checklist item does not specifically require that the bacteria be detected by culturing the platelets. Second, the checklist item is phase I (not phase 2). Third, even with a method as insensitive as dipstick or Gram stain, one may detect a massively contaminated unit that should not be transfused Only leukocyte-reduced products can be tested by the commercial culture systems available (Pall BDS and BioMerieux BacT/ALERT). The FDA has approved both Pall BDS and BioMerieux BacT/ALERT for quality control testing (but not bacterial screening) of leukocyte-reduced platelets. In the case of the Pall BDS, which is based on the demonstration of a reduction in oxygen, it is important that the sample submitted for testing be free of metabolically active cells that may utilize oxygen and result in a false-positive result. This same rationale may be true for the BioMerieux BacT/ALERT, which detects the generation of carbon dioxide rather than the reduction of oxygen. In addition, there are alternatives, albeit less automated and/or sensitive than desired. Dipsticks and Gram stains can be implemented and may detect significantly contaminated units. Quantitative bacterial plate cultures may also be used. 5 Other alternative solutions are in sight If one uses BacT/ALERT and given it requires several days to report results, will there be product recalls after transfusion, resulting in patient anxiety and considerable needless effort? Is the false-positive rate known for this method? The BacT/ALERT does not require several days to generate a report, and available data indicate that most positive results for pathogens are detected in the first 12 to 24 hours following incubation. Nonetheless, since current practice with that test system is to extend culture for the life of the platelet product, this testing strategy may delay the reporting of positive results. Because culture results will occasionally become positive after a unit of platelets has been distributed for use, product recalls are likely to occur, and in some instances a recalled platelet unit will already have been dispensed for transfusion. In the event a recalled unit has been dispensed for transfusion, the recipient s physician and/or health care providers will need to be notified immediately. Fortunately, recalled units that turn positive by BacT/ALERT after 24 hours of culture usually harbor low-level bacterial contamination and/or slow-growing bacteria with limited pathogenic significance. The false-positive rate as reported with the BacT/ALERT is very low, in the range of only 0.02% to 0.04%. On the other hand, Brecher et al 17 reported that quality control testing with BacT/ALERT on 2397 apheresis platelets collected during an 11- Arch Pathol Lab Med Vol 128, September 2004 Shulman 961

4 month period, showed an overall true-positive rate of 0.29%, with a true-positive rate for aerobic organisms of 0.13% and an anaerobic truepositive rate of 0.17%. 8. New standards were instituted without discussion among the entire transfusion medicine community; consensus is needed. The TMRC respectfully disagrees. There have been extensive discussions on this issue with numerous opportunities for input from the entire transfusion medicine community. The problem was discussed in the early 1990s in a JAMA article entitled Septic Reactions to Platelet Transfusions: A Persistent Problem. 4 The problem was part of the 1980s discussion regarding 5- versus 7-day storage for platelets. The FDA has convened 3 national conferences on this issue. Consensus has in large part been achieved, even if unanimity has not. The following is an opinion piece by Roger Dodd, written when he was AABB president during (Reprinted with permission from AABB News [May/June 2003]. Copyright 2003, American Association of Blood Banks.) In March 2003, AABB s Board of Directors approved the 22nd edition of Standards for Blood Banks and Transfusion Services. Over the years, the standards have introduced many required activities some more welcome than others but all aimed at improving transfusion safety for patients. As a board and as an organization, we have received extensive feedback about many of these initiatives from you, the members. In the past couple of years, the most common lament has been the inability of the blood banking community to prioritize, by importance, threats to transfusion safety. The new edition of Standards presents the transfusion and blood banking community with a unique opportunity, that of responding to an issue identified as one of the greatest threats to transfusion safety. Unlike other challenges, the threat of bacterial contamination has been an issue owned and highlighted by our own membership. One year ago, the AABB Board asked two prestigious committees, the Transfusion Transmitted Diseases Committee and the Clinical Transfusion Medicine Committee, to assess current risks to the blood supply. Members of these committees, experts in blood banking and transfusion medicine from all sectors of our profession, concluded that the No. 1 risk to infection-related transfusion recipients is bacterial contamination. Since that time, the consensus of the blood community over the need to address bacterial contamination has been growing. In 2003, members of our own medical discipline wrote an open letter to the blood community, urging them to address the risk in a meaningful way. In January, the Advisory Committee on Blood Safety and Availability adopted a resolution, calling on the federal government to take action to detect bacterial contamination in blood components. The College of American Pathologists recently incorporated a requirement in its Transfusion Medicine Checklist, requiring that laboratories have a system to detect the presence of bacteria in platelet components. Unlike during other threats to blood safety, existing or theoretical, that have been hyped by the media and the public, the groundswell for reducing bacterial contamination has come from the blood banking and transfusion medicine community. The Standards Committee listened to our profession and took action. Unfortunately, the solution to the problem will prove to be as difficult if not more so than identifying the problem. There is no clear path to eliminating the risk; there is no unanimity about how the standard should be written to effect the change. There are regulatory hurdles to be addressed if the solutions are to be effective, and there are unbudgeted costs associated with the fix. But these obstacles are ours to address, without the constraints of regulatory timelines and public pressure to act. Our most commonly voiced complaint today is that we are forced to operate under an agenda that we do not control. Bacterial detection presents us with an opportunity to control our own destiny, to demonstrate that we can and do lead the profession, and that our paramount concern is to improve transfusion safety for patients. It is an opportunity that we cannot allow to pass. 9. One could argue implementation should await FDA acceptance of an extended platelet shelf life standard and pooling of random donor platelets with testing. One could argue this, but FDA acceptance of 7-day platelet storage will most likely need to await FDA approval of systems for extended storage coupled with screening for bacterial contamination using a method cleared by the FDA for bacterial screening. It is likely that studies of sufficient size will be required to satisfy the FDA and that completion of these studies will be slow. 10. The development of a new standard was influenced by proponents with a vested interest, such as a research focus in this area and/or funding derived from manufacturers of bacterial detection systems. How much of a factor is this? Is it a factor at all? This is an interesting concern, suggesting that undue conflicts of interest are driving the field toward bacterial detection testing of platelets. A discussion of this issue has been posted for public view on the Internet. 19 The TMRC members believe that such concerns are reminiscent of similar discussions that blood bankers had regarding implementation of radioimmunoassay, enzyme immunoassay, nucleic acid amplification testing, and other technologies that have markedly improved the safety of the blood supply within the United States. The members of the TMRC have come to the consensus opinion that TRM should be included in the CAP LAP checklist for the sake of the patient. Being a researcher in the area of bacterial contamination of platelets, or being a consultant to a company that manufactures equipment or test kits for detection of bacteria, does not imply complicity. In summary, it remains the consensus view of the CAP TMRC that TRM is an appropriate item for inclusion on the CAP LAP checklist. Dr Shulman is advisor to and immediate past chairman of the CAP Transfusion Medicine Resource Committee. He is also a member of the Medical Advisory Board of Verax Biomedical, which is located in Worcester, Mass ( veraxbiomedical.com/subpage.asp?id 4&nid 1). References 1. Ness P, Braine H, King K, et al. Single donor platelets reduce the risk of septic platelet transfusion reactions. Transfusion. 2001;41: Perez P, Salmi LR, Follea G, et al. Determinants of transfusion associated bacterial contamination: results of the French BACTHEM Case-Control Study. Transfusion. 2001;41: Kuehnert MJ, Roth VR, Haley NR, et al. Transfusion transmitted bacterial infection in the United States, 1998 through Transfusion. 2001;41: Morrow JF, Braine HG, Kickler TS, Ness PM, Dick JD, Fuller AK. Septic reactions to platelet transfusions: a persistent problem. JAMA. 1991;266: Yomtovian R, Lazarus HM, Goodnough LT, Hirschler NV, Morrissey AM, Jacobs MR. A prospective microbiologic surveillance program to detect and prevent the transfusion of bacterially contaminated platelets. Transfusion. 1993;33: Hillyer CD, Josephson CD, Blajchman MA, Vostal JG, Epstein JS, Goodman JL. Bacterial con- 962 Arch Pathol Lab Med Vol 128, September 2004 Shulman

5 tamination of blood components: risks, strategies, and regulation: joint ASH and AABB educational session in transfusion medicine. Hematology (Am Soc Hematol Educ Program). 2003: Chiu EK, Yuen KY, Lie AK, et al. A prospective study of symptomatic bacteremia following platelet transfusion and of its management. Transfusion. 1994;34: Noel L, Debeir J, Cosson A. The French haemovigilance system. Vox Sang. 1998;74(suppl 2): Debeir J, Noel L, Aullen J, et al. The French haemovigilance system. Vox Sang. 1999;77: Williamson L, Cohen H, Love E, Jones H, Todd A, Soldan K. The Serious Hazards of Transfusion (SHOT) initiative: the UK approach to haemovigilance. Vox Sang. 2000;78(suppl 2): Burns KH, Werch JB. Bacterial contamination of platelet units: a case report and literature survey with review of upcoming AABB requirements. Arch Pathol Lab Med. 2004;128: Commission on Laboratory Accreditation. Laboratory Accreditation Program transfusion medicine checklist: December Available at: accreditation/ checklists / transfusion medicine December2002.pdf. 13. Jafari M, Forsberg J, Gilcher RO, et al. Salmonella sepsis caused by a platelet transfusion from a donor with a pet snake. N Engl J Med. 2002;347: Yomtovian R. Bacterial contamination of blood: lessons from the past and road map for the future. Transfusion. 2004;44: Use of 7-day platelet storage in the United States: letter to United States Customers: October 30, Available at: download.asp?area public&file A.pdf. 16. Hemosystem: rapid detection of rare events in blood products. Available at: hemosystem.com. 17. Brecher ME, Hay SN, Rothenberg SJ. Monitoring of apheresis platelet bacterial contamination with an automated liquid culture system: a university experience. Transfusion. 2003;43: Dodd RY. President s message: our voice at work. AABB News. May/June 2003: Transfusion medicine leaders urge blood collection community to immediately initiate program to detect bacteria in platelet products [CBBS e-network Forum, posted August 29, 2002; addenda posted August 30, September 4, 6, 11, 13, and 16, 2002]. Available at: Arch Pathol Lab Med Vol 128, September 2004 Shulman 963