Blood safety: bacterial screening of blood products
Erica Wood, Australian Red Cross Blood Service, PO Box 354, 3205 South Melbourne, Victoria, Australia
Introduction: what is the problem?
Bacterial contamination of blood components can cause death and serious morbidity in transfusion recipients [1–3]. Numerous measures to prevent and detect bacterial contamination have been introduced to reduce this hazard. Detection methods, including screening of blood components, are the subject of this review. Preventive measures include careful donor selection and predonation health screening, with donor deferral and follow-up of postdonation information as required [4,5]; optimization of blood collection and manufacturing processes, including improvements in skin cleansing procedures [6–9], choice of collection method and consumables, diversion of initial volume away from the collection bag [10–16], environmental monitoring, and greater availability of the use of apheresis platelets to minimize exposure to donor skin flora [17,18]; developments in pathogen reduction technologies [19–24]; attention to component storage and handling; and promotion of evidence-based clinical transfusion practice [24,25].
However, in spite of the introduction of advancements such as these, haemovigilance programs internationally continue to report serious adverse events because of bacterial contamination. For example, in spite of an overall reduction in numbers of cases reported annually to the French national haemovigilance programme since 2000, there were nine clinical cases in 2008 with imputability ratings of ‘likely/possible’ to ‘certain’, including one death because of Escherichia coli contamination of a platelet transfusion . In the same year, the annual report from the Serious Hazards of Transfusion (SHOT) scheme described six confirmed clinical cases of sepsis in the United Kingdom because of bacterial contamination of platelet units, with consequences ranging from minor morbidity to death .
The spectrum of bacterial contamination ranges from detection of bacteria on routine component surveillance testing without apparent consequence to fatal sepsis. Recognition of transfusion-related sepsis can be difficult at the bedside, especially where patients may be already febrile and/or on antibiotics for suspected sepsis from other sources [28,29]. In order for bacterial contamination events to be captured by haemovigilance systems, the cases must first be identified clinically, confirmed with appropriate investigations, and then reported at institutional and programme levels.
Haemovigilance programs are at various stages of development in different regions. Where they do exist, in some jurisdictions participation is mandatory and in other cases voluntary and therefore not universal. Different systems require variable levels of clinical certainty and laboratory confirmation, and some cases which are deemed highly likely on clinical grounds are unable to be confirmed because of unavailability or inappropriate handling or storage of blood component or patient samples. Thus, the number of confirmed and reported transfusion-related clinical cases can be considered the ‘tip of the iceberg’ or perhaps ‘the nose of the crocodile’– with the majority of bacterial contaminations of blood components hidden below the surface, and many associated with no apparent harm, but with the potential for devastating outcomes. The cases included in haemovigilance reports therefore represent only a subset of total bacterial events.
Even the definition of what constitutes a case of transfusion-transmitted bacterial contamination is a matter of some discussion. SHOT  defines a confirmed transfusion-transmitted infection 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.
The Dutch Transfusion Reactions in Patients (TRIP) haemovigilance system has recently expanded its scope of reportable bacterial events . To the previous category of ‘post-transfusion bacteraemia/sepsis’, defined as ‘clinical symptoms of bacteraemia/sepsis arising during, directly after or some time subsequent to a blood transfusion, for which there is a relevant, positive blood culture of the patient with or without a causal relation to the administered blood component’, the following types of cases are now captured by TRIP:
- • ‘Bacterial contamination of blood component’ events, where ‘…..bacteria in a (remnant of) blood component or in the bacterial screen bottle of a platelet component, or in material from the same donation, demonstrated in the approved way with laboratory techniques, preferably including typing of the bacterial strain or strains’, and
- • ‘Positive bacterial screen’ events, where ‘the blood service reports a positive bacteriological screen, but bacterial contamination of the relevant material is not confirmed by a positive culture result on the same material or other products made from the same donation.’
By expanding the scope of reportable events, TRIP hopes to better understand the spectrum of clinical scenarios in transfused patients, including ‘a better picture of the influence that transfusion and bacterial contamination/infection could have on one another’. Further description of these interesting developments is included in the programme’s annual report. Consensus definitions for bacterial events are being developed jointly by the ISBT Working Parties on transfusion-transmitted infectious diseases and haemovigilance.
Detection of bacterial contamination of blood components
Strategies to detect the presence of bacteria in blood components include:
- • quality control testing
- • routine surveillance screening or other prerelease testing
- • visual inspection of components prior to issue from the blood centre, at the hospital transfusion laboratory, and at the bedside prior to administration
- • careful monitoring of patients during and after transfusion
Screening blood components for bacteria
Routine prerelease surveillance screening of platelets is now in place in a number of countries. It is based on the rationale that platelets are the blood components most at risk because of the current requirement for room temperature storage, and that while the presence of detectable bacteria is not always associated with clinical consequences, the detection of bacteria using sensitive methods may prevent issue and transfusion of at least some contaminated units [31,32]. In some countries, use of platelet surveillance cultures has been used to extend shelf-life beyond 5 days, either routinely or occasionally .
The challenges associated with platelet screening are substantial [2,31,33–35]. These include some biological characteristics of the contaminating organisms, such as an initial very low inoculum at the time of blood collection and bacterial kinetics, with an initial lag phase prior to exponential proliferation; some logistical aspects of preparation and supply for a blood component with a short shelf-life; and some aspects relating to the testing methods. Many organisms will take up to 24–48 h to grow in the culture-based systems currently in widespread use.
Measurement of pH and glucose, or evaluation for swirling, is simple and inexpensive, but not sufficiently sensitive or specific for routine use in detection of bacterial contamination [36–38].
Experience from culture-based surveillance programs has now been reported by institutions from a number of countries over some years [39–62]. The most widely used of these are the BacT Alert®3D system (bioMérieux, Durham, NC, USA) and the Pall eBDS (Pall Corporation, East Hills, NY, USA). Each has been shown to be suitable for culture of red cell and platelet components, in blood centre and hospital settings. However, comparison of findings, and therefore determination of the optimal combination of sample and method, can be difficult, because of:
- • differences in types of components tested (apheresis or whole blood-derived platelets, method of manufacture, whether in additive solution or plasma)
- • time from collection to sampling (3–36 h)
- • volume inoculated (3.5–10 ml per component for each culture bottle)
- • type of testing system, culture vial and detection method
- • duration of culture following sampling and prior to release (0–24 h)
- • total duration of culture (5–7 days)
- • use of aerobic culture alone, or combined with anaerobic culture
- • definitions of true and false positive, confirmed and unconfirmed results
- • availability of confirmatory testing and
- • active or passive follow up of transfused units and clinical cases
It remains debatable whether the routine addition of anaerobic culture to aerobic cultures is helpful in detecting additional bacterial contamination events. On the one hand, screening systems using aerobic plus anaerobic cultures typically identify more organisms than those using aerobic culture only, and strict anaerobes may rarely be associated with serious clinical outcomes. Of the transfusion fatalities reported to the US Food and Drug Administration since 2005 because of bacterial contamination, including five cases because of platelets in 2009 alone, the vast majority were facultative anaerobes, and one was because of a strict anaerobe (Eubacterium limosum) . However, many of the organisms identified only by anaerobic culture are not considered clinically significant, and a higher rate of detection may be obtained simply by increasing the sampling volume into aerobic culture bottles [55,59], so the role of routine anaerobic culture as a way of increasing detection of meaningful bacterial contamination remains a topic of discussion.
In Australia, routine screening was introduced for all platelets in April 2008. This involves a 15–20 ml sample obtained using closed system sampling at 24 h, with 7 ml inoculated into each of an aerobic and anaerobic culture bottle and cultured using the BacT Alert system (bioMérieux). Platelets are released as ‘negative to date’. A national recall office notifies hospitals if the machine flags an ‘initial machine positive’ after a component has been released, and a member of the transfusion medicine team follows up positive results with treating clinicians where components have been transfused. While the majority of organisms identified have been corynebacterium or propionibacterium species, over 40 organisms considered potentially clinically significant have been identified and prevented from transfusion, including Klebsiella, Serratia and Clostridium species. Where platelets have been transfused, units which are ‘initial machine positive’ appear to have low risk for clinical consequences. No cases of confirmed or high-probability transfusion-related sepsis have been reported to the Australian Red Cross Blood Service since screening was introduced .
However, false-negative results can occur so bacterial screening does not eliminate all risk. Cases have been reported of clinically significant events following transfusion of platelets with negative screening results [44,51,52,56,57,61], so vigilance must be maintained even where multiple measures to prevent and detect bacteria are in place, and bacterial contamination of red cells remains an important, potentially fatal event, even if it is uncommon.
Monitoring the effect of these interventions also requires balancing the reduction in bacterial contamination achieved through screening against the fact that components are typically slightly older at time of release from the blood centre (because of the ‘hold’ period prior to sampling, plus the time taken in the testing process) as well as the requirement to maintain sufficient inventory of these short shelf-life components.
Other bacterial screening methods
Several rapid detection methods for use immediately prior to transfusion are available, and others are in development. Flow cytometry appears somewhat less sensitive than culture-based methods but may be suitable for testing platelets prior to issue for transfusion. It is relatively simple and fast to perform, and uses equipment and techniques already widely available in many laboratories [65–67]. Molecular methods such as real-time PCR show high sensitivity and specificity, with results available within hours [68–70]. An immunoassay (Platelet PGD, Verax Biomedical) is already available and detects antigens common to Gram negative and positive bacteria. The system overall appears somewhat less sensitive than other methods currently available, and it is likely to be best suited for testing of a small volume sample in the immediate pretransfusion setting, as it is simple and results are quickly available. Another method showing potential is a non-invasive near infrared system which has been described recently . Most of these methods have been compared directly to the ‘gold standard’ of culture-based systems, and several newer methods have been compared to each other [72–74], but further work remains to be carried out to understand how and when they might give the most useful results and to optimize them for routine use in the blood centre and/or hospital environment.
Availability of suitable standards for blood component testing will assist comparisons of laboratory data from bacterial screening programs, and in evaluation of new technologies. The ISBT Working Party on transfusion-transmitted infections has recently co-ordinated an international collaborative validation study of candidate standards.
Conclusions – and some of the many important remaining questions
Much has been achieved in our efforts to improve the bacterial safety of transfusion, and undoubtedly lives have been saved by the interventions already introduced, yet important challenges and questions remain. Most published reports on bacterial prevention or detection measures relate to laboratory measurements of outcome – but how do these translate into reductions in clinical risk, and in particular, to serious adverse events and fatalities? Diversion pouches used in collection systems appear to greatly reduce the chances of bacterial contamination of platelets with skin flora, but how do we address the residual risks from other organisms, and how best to lessen the possibility of red cell contamination? What proportion of culture-positive units are associated with clinical consequences? How do we reduce the substantial rates of false positive, indeterminate and unconfirmed initial results from screening programs, which have both clinical and product wastage implications? Are molecular studies necessary to confirm identity between bacteria detected in a blood component and in a patient sample, and if so, is this a requirement for definition of a clinical case? What is the place of pathogen inactivation systems in reducing the risk of bacterial contamination of blood components, when some of these systems do not inactivate spores, and if pathogen reduction was routinely in place, would other measures still be required? These and many other questions remain to be addressed.
The combination of efforts to reduce bacterial contamination of blood components [1,31,34,35,59,75], of which screening is just one element, must also be matched by efforts to promote more appropriate and evidence-based clinical practice, to improve awareness and management of transfusion-related sepsis and reporting of adverse events to bring real benefits for patients.
Dr Marija Borosak and Dr Janet Wong provided helpful data and analysis of results and clinical follow-up of bacterial screening results from the Australian Red Cross Blood Service. Dr Frank Hong kindly reviewed the draft manuscript.
No conflicts declared.