Prevention of bacterial contamination, including initial flow diversion

Blood centers and hospitals worldwide face a continuing challenge to improve the safety of blood components for transfusion. Important improvements have been achieved due to careful donor selection, and advances in collection, testing, storage and handling. Transfusion-transmitted infections (TTIs) are now rare in many countries. However, haemovigilance system reports reveal that bacterial contamination of blood components remains an important residual risk, particularly relating to platelet transfusions. For example, the 2008 annual report from the Serious Hazards of Transfusion (SHOT) scheme in the United Kingdom describes a total of 38 confirmed TT bacterial events since 1998 [1]. The majority (32) of these were related to platelets and the others (six) to red cells. Forty patients were affected in total, as some apheresis platelets were split and went to different recipients, with both patients suffering an adverse event. In 2008, all four confirmed cases were due to platelets, with consequences ranging from minor morbidity to death in a total of six patients. This can be contrasted with 22 confirmed TT viral or parasitic infections in 25 patients reported to SHOT since 1998, and no confirmed infections reported between 2006 and 2008. Clearly, therefore, this is an important area where ongoing efforts are warranted to improve transfusion safety. This review will focus on efforts to prevent bacterial contamination occurring, with some consideration of methods to detect it, where present, in order to prevent contaminated units reaching the patient. In some clinical cases the presentation of bacterial contamination is readily recognized. However, in many situations confirmation of a clinically-suspected TT bacterial event is not easy, and the ratio of confirmed to suspected cases described in haemovigilance reports is low [2]. This relates partly to the recipient population, especially for platelet transfusions, where many are transfused to hematology-oncology patients [3, 4] who are often already immunocompromised, whether by underlying disease, or effects or complications of therapy, such as neutropenia. These patients may already be febrile and/or on broad-spectrum antibiotics [5, 6], and typically they have indwelling catheters which may also be a source of bacteremia. Therefore, post-transfusion fever in these patients may not be appreciated as a sign of a bacterially contaminated component. However, in a prospective study, Chiu and colleagues [7] reported that in neutropenic transplant patients experiencing fever following platelet transfusions, especially where the fever rose 2°C or more from baseline, these events were frequently associated with bacteremia, and sometimes serious complications, including septic shock. The diagnostic challenge also reflects the diverse clinical presentations, which can range from no signs whatsoever to fever, hypotension and shock [5, 8, 9]; even where these are present it can sometimes be difficult, at least initially, to separate a bacterial event from other potential causes of serious transfusion complications such as acute hemolytic reactions. However, a high index of suspicion is essential if these cases are to be recognized promptly and managed appropriately. SHOT [1] defines a confirmed TTI as a case where: The recipient had evidence of infection post-transfusion, and there was no evidence of infection prior to transfusion, and no evidence of an alternative source of infection; and, either at least one component received by the infected recipient was donated by a donor who had evidence of the same transmissible infection or at least one component received by the infected recipient was shown to contain the agent of infection”. A number of other haemovigilance programs, including the regional Serious Transfusion Incident Reporting (STIR) system in Australia, have adopted definitions based largely on the SHOT definition [10]. Consensus definitions for bacterial contamination events are being developed jointly by the ISBT Working Parties on TTI and Haemovigilance. Not infrequently, implicated components are either not available for examination, having been discarded, or their storage or handling prior to laboratory testing will have compromised the results. Therefore, in some situations even where the clinical picture strongly suggests a bacterial event, the case may not be confirmed. In these situations, it is still important that the case be reported and investigated to the extent possible, with imputability assigned as appropriate (possible, probable, unconfirmed etc). Consideration of an additional definition of bacterial contamination may be required, especially since the recent introduction of routine surveillance screening of platelets in many countries. For example, the Transfusion Reactions in Patients (TRIP) haemovigilance system in The Netherlands has introduced a new category for reporting positive results of surveillance screening (bacterial contamination of blood component), in addition to the existing classification for clinically apparent transfusion reactions (post-transfusion bacteremia) [2]. This review does not permit a comprehensive analysis of the microbiological aspects of bacterial contamination. The interested reader is referred to other sources for more information [11-14]. However, suffice to say here that a wide range of bacteria have been reported in association with contamination of blood components from clinical case series, haemovigilance reports, detection by quality control testing and more recently from the introduction in some countries of routine bacterial surveillance screening of platelets. Many of these organisms are able to grow up during component storage, and this is especially associated with the growth of skin-derived flora during room temperature storage of platelets. Several reports have demonstrated an association between duration of platelet storage and likelihood and degree of bacterial contamination [1, 9, 15]. Other sources of contamination include donor bacteremia [16, 17] and contamination during collection, manufacturing or subsequent handling [18, 19]. Of these, many but not all appear to cause clinically apparent sepsis in transfusion recipients. Gram negative bacteria, especially where associated with endotoxin production, are associated with a substantial proportion of severe cases and fatalities [8, 14]. Some Gram negative organisms such as Pseudomonas and Serratia species grow well at the refrigerated temperatures of red cell storage. However, even seemingly innocuous bacteria such as common skin flora may be associated with very serious outcomes, especially at high levels of contamination. On the other hand, the clinical consequences of receiving a bacterially contaminated component are not always serious. Reports from blood services and haemovigilance systems in countries where routine screening of platelets is now in place describe a low rate of clinical events when patients receive platelets issued as ‘negative to date’ but where screening results result in clinical follow up. This is especially so where these are either unconfirmed (no growth after an “initial machine positive” signal, usually a system-related event), indeterminate, or found to contain organisms such as Propionibacterium acnes [20-22]. This may in part reflect the pathogenicity of some of the organisms likely to be detected late during platelet culture, such as P acnes. Even where organisms considered more likely to be clinically harmful are present in the component, their rapid growth in many cases appears to correlate with the ability to detect them and remove them from inventory prior to transfusion, although this is not always so. In other situations, patients may be protected if already on antibiotics with coverage against the bacteria detected in the component. Important elements in reduction of bacterial risk include careful blood donor selection, based on relevant health history and examination, to exclude those with identified risks for bacterial contamination of blood components. Most blood centers have well established policies and procedures regarding donor bacterial risks, however there is relatively little published literature on the effectiveness of the questions asked prior to donation. While undoubtedly helpful to prompt donors to consider symptoms suggesting potential bacterial risks, such as recent dental procedures, skin lesions and gastrointestinal illness, the sensitivity and specificity for bacterial risks is not known [23, 24]. Donors should also be encouraged to report relevant postdonation information promptly, to enable quarantine of at risk units, to prevent transfusion and also to permit component testing where appropriate. “. . . Complete sterilization of the skin previous to phlebotomy is impossible, since there are always a few bacteria which escape the bactericidal action of any cutaneous antiseptic . . .” Novak (1939), cited by Yomtovian [25]. Careful attention by collection staff to phlebotomy site preparation is very important. Most bacteria identified as contaminating platelets are typical skin flora, but several illustrative cases have been reported due to unusual organisms, such as Salmonella and Clostridium perfringens [26, 27]. While even the best skin cleansing regimens will not remove all bacteria present on the skin or in its appendages, improved skin disinfection measures have been reported to reduce presence of bacteria in blood components. A number of groups have described good results with the use of 70% isopropyl alcohol [28], various iodine preparations, or 1% chlorhexidine [29], used alone or as combinations in multi-step procedures [30, 31]. For more complete information the reader is referred to the review by McDonald [32]. A recent Cochrane review [33] summarized available trials of skin disinfection in the blood donation context and other settings. However, because of the many methodological differences, such as randomised or other design, plus differences in study population (donors or patients), type of agent used, one- or two-step procedures, method and timing of application, and different outcome measures (whole blood or final component contamination rate by culture, or clinical events), the Cochrane reviewers were unable draw any firm conclusions regarding differences between various methods and their effectiveness in reducing bacterial contamination of blood components. Diversion of the initial volume of blood away from the collection bag has been introduced in many countries following studies showing reduction in bacterial contamination of both whole blood and various components [28, 34-36] and as expected this appears especially effective at reducing skin flora [37]. Diversion volumes are most commonly reported as between 15 and 25 ml, with volumes described from 10 ml to 50 ml. These samples can be used for other purposes, such as blood group and viral testing. Satake and colleagues in Japan demonstrated a reduction of 71% when comparing cultures of expired platelets (at least 4 days old) before and after introduction in 2006 of a 25 ml diversion volume [38]. This is comparable with the experience from other international groups reporting reduction of between 40 and 88% on bacterial testing of either whole blood or components. This area has recently been reviewed by Liumbruno and colleagues [39] and other authors have described their experiences in the combined effects of introduction of both improved skin disinfection and diversion methods [28, 35, 40, 41]. Observance of the principles of Good Manufacturing Practice (GMP) is essential for the preparation and supply of safe blood components. Critical elements are the availability of staff trained and competent in their tasks, and adherence to documented standard operating procedures (SOPs) for donor assessment, blood collection, and processing, storage and handling. A description of bacterial contamination events due to contaminated chlorhexidine used in preparation of the phlebotomy site emphasizes the importance of vigilance in all aspects of the process [42]. Choice of collection method (for example, one- or two-arm procedure for apheresis platelet collections) should be carefully considered in the operational context of other measures to reduce bacterial contamination [43]. Appropriate equipment and consumables, such as blood collection kits to permit closed system component separation and other manipulations, such as leukocyte depletion, is very important. Proper cleaning and maintenance of the facility and equipment should be in place, with environmental monitoring as required. Correct storage of whole blood and components during transport to and from the blood center and hospital, and during preparation, must be maintained, including meeting time and temperature specifications at each stage of the process. The ideal bacterial detection system for blood components would be simple, quick to perform either at the blood centre, or at the bedside, reliable and robust, inexpensive, highly sensitive (to detect all relevant cases), specific (to prevent unnecessary work and component discard) and give rapid results [44]. Not much to ask, really! The ideal system does not yet exist, but a great deal of work has occurred in this area in recent years. Existing technologies are being refined, and new techniques are becoming available. Detection methods, used immediately prior to component issue or administration, may be able to be less sensitive than those employed very early to detect the presence of bacteria. The most widely used culture-based methods are the BacT Alert®3D system (bioMérieux, Durham, NC) and the Pall eBDS (Pall Corporation, East Hills, NY). Each has been shown to be suitable for culture of red cell and platelet components, in blood center and hospital settings [9, 45-47]. In recent years, some centers have introduced routine bacterial screening for platelets using one of these methods. They both have very high sensitivity [21, 45, 46, 48] and are designed to detect small numbers of bacteria present in the initial inoculum, but also often have substantial rates of initial reactive results which are later not confirmed [49-51]. This can result in unnecessary discard of components and substantial additional work for blood services and hospitals in following up components which have been issued for transfusion, as well as anxiety for patients and clinicians. Differences in method and timing of sampling, volume inoculated, type of container into which samples are inoculated, use of aerobic culture alone or with anerobic culture, type of component and method of preparation, quarantine period of components after hold and prior to issue, and definitions of confirmed positive results, among other factors, can make comparisons between published reports somewhat difficult [9, 28, 41, 43, 50, 52-57]. The ideal volume and timing of sampling is not established, although models exist to predict yield of the combination of testing elements [58, 59]. With current systems, small volume and earlier sampling and issue risks missing bacteria present at very low levels after collection, due to sampling error related to initial low bacterial load and the lag phase for bacteria to enter the phase of logarithmic proliferation. This must be balanced against the need to issue platelets as fresh as possible to deliver maximum clinical benefit, yet there is a greater chance of detecting presence of bacteria when sampling is performed after a longer hold and where platelets are issued after the culture is ‘negative to date’ for a longer period. These complexities are considered in a review by McDonald [32]. In Australia 15–20 ml is removed, using a closed system, from each platelet component at 24 h after collection, and samples are inoculated for aerobic and anaerobic culture. Platelets are released, typically on day 2, as ‘negative to date’ while the culture continues over the component shelf-life. In some (predominantly European) settings but not others (for example, New Zealand [60], Australia, Hong Kong SAR), use of platelet surveillance cultures has been used to extend shelf-life beyond 5 days, either routinely [22, 50] or occasionally, either as part of a research protocol [61] or on clinical grounds [54], and sometimes on the basis of a re-test after initial sampling (Foley et al., in [56]). In Japan, platelet shelf life remains at 72 h [38]. However, screening programs do not always detect bacterially contaminated components for a range of reasons, and false negative screening results do occur. Cases have been reported of clinically significant events following transfusion of platelets with negative screening results [20, 43, 55, where these surveillance systems have been introduced they are considered to have to a reduction in bacterial risk, with a number of programs recently reporting their experience of of the majority of components in which clinically significant bacteria were detected in time to prevent transfusion [20, 41, Some rapid for use immediately prior to transfusion are available and others are in cell analysis has been for use with platelet components. appears somewhat less sensitive than culture-based methods but may be suitable for testing prior to issue for transfusion, being relatively and to and using equipment and techniques already widely available in many methods such as have been reported to high sensitivity and with results available than The reader is referred to recent for more information and colleagues rapid detection and and found to be the most sensitive but with good sensitivity of all methods. should be that the is no longer available. of the other methods is already available. common to Gram negative and positive bacteria. The system appears less sensitive than other methods but it is likely to be best for testing of a small volume in the being to results are available Other methods and or for while to perform and inexpensive, are not sensitive or specific for routine use in detection of bacterial contamination for is not at all sensitive but remains very important prior to administration, as this is the final chance to prevent of a contaminated component. reduction technologies to platelets already These include and and and is most experience with the system, which has been used for several and is now in place in a number of and others are in various of but are not yet in routine clinical Other methods such as use of are in The is to platelet with and least potential substantial work has been in for red this is less and will not be of for platelets include the ability to a wide range of bacteria and other early after collection, prompt issue for transfusion and the of shelf-life from 5 to include of bacteria, and some of platelets during In some studies this into platelet transfusion for patients In report there was no in platelet these studies used different methods to platelet use While use of platelets appears to be possible effects of residual levels of components such as especially in patients who may have over and even years, are not yet remains to be The reader is also referred to recent and the which the important questions and in this The ISBT Working on an international study of blood bacteria during 2008. The study was performed in to the of bacterial the species and relevant to the of blood components. were by the and to results were to the Working at the ISBT regional meeting in early and have since been to the for for The study results and the availability of suitable will comparisons between and between methods, including in of new technologies for detection and reduction of bacteria in blood components. Most bacterial contamination events presence of bacteria in the transfused component. In of one to has been to donor and colleagues reported sepsis as more than from platelets from six donor in their than with apheresis components donor They also described that as the proportion of apheresis platelets over time to in their bacterial contamination events were On the other hand, more recent reports from platelet surveillance screening studies [28, found no significant between rates of confirmed bacterial contamination in and apheresis platelet components. This may reflect of improved skin disinfection methods and diversion techniques in recent years, although other may also including differences in type of bacteria detected by surveillance screening with those reported from clinical case While some recent reports have described more reactions with apheresis platelets, this appears to reflect a much greater use of apheresis components in to platelets, for a range of clinical and The haemovigilance found no in clinical rates between the of components. While the differences in bacterial contamination rates between platelet component appear to have some authors still apheresis platelets as one to reduce to bacteria and apheresis platelets to other important of donor such as potential for and other Transfusion is widely but often not Many have transfusion policies and procedures, including relating to and of transfusion yet are frequently not of and staff have often had little in clinical transfusion and are not risks, including the small but potential for bacterial contamination of blood components, or in which these can be including storage and and component prior to A recent study of found little in this or experience of a transfusion due to bacterial contamination A described substantial in suggesting the need for availability of and to transfusion appropriate storage and careful prompt of signs of and appropriate can all remains yet to be in this area of clinical transfusion A combination of appears essential to the important residual risks of bacterial contamination of blood components. are good laboratory showing reduction of bacteria in components in to improved skin disinfection techniques and diversion of initial collection and the of quality control The of surveillance screening programs in transfusion of contaminated components has already been demonstrated in some settings. all to these efforts now into clinical which will be in clinical case and haemovigilance McDonald and helpful on the