Bacterial detection and extended platelet storage: The next step forward

Platelet (PLT) concentrates are stored under aerobic conditions at room temperature, circumstances that are ideal for bacterial growth. Large multicenter studies with licensed bacterial detection systems now confirm a significant rate of bacterial contamination that constitutes a major risk to patient safety. The application of AABB standard 5.1.5.1 to apheresis PLTs removed more than 100 contaminated units from the US blood supply in its first year and has almost certainly prevented both septic reactions and fatalities.1-3 This success has come at a cost of a 3 to 5 percent price increase, 1 to 2 percent product loss due to sampling, a 12- to 30-hour delay in product release resulting in an effective product shelf life in the transfusion service of 3 to 3.5 days, and reduced PLT availability as demand for apheresis PLTs increases owing to the lack of a practical licensed test for individual whole blood–derived (WBD) PLTs. Bacterial testing, however, provides an opportunity for an extension of PLT shelf life from 5 to 7 days that will help hospitals manage their inventories and reduce outdates.4 The Food and Drug Administration (FDA) has given Gambro BCT (Lakewood, CO) approval to implement a large postmarket surveillance study of extended storage PLTs, termed the “Post Approval Surveillance Study of Platelet Outcomes, Release Tested” (PASSPORT).3 As this study is implemented and other companies propose their own protocols, it is likely that extended-shelf-life PLTs will become widely available. At this point it is worth reflecting on what we have learned from bacterial screening and to ask whether the cost of extended storage with current bacterial testing technologies is worth its presumed benefits. Three testing systems are licensed in the United States for PLT quality control (QC): the BacT/ALERT 3D automated microbial detection system (bioMérieux, Durham, NC); the enhanced bacterial detection system (eBDS; Pall Corp., East Hills, NY); and the ScanSystem (Hemosystem, Marseille, France). Extensive data are now available on the use of the BacT/ALERT and eBDS systems in routine blood center operations. te Boekhorst and coworkers5 report the use of the BacT/ALERT system to test pools of five buffy-coat PLT units. Ten to twenty milliliters of product was harvested 14 to 24 hours after collection and cultured under both aerobic and anaerobic conditions, while the products were released into inventory. They demonstrate an initial positive rate of 0.7 percent (203 of 28,104 pools), leading to the recall of the implicated products and their associated red blood cell and plasma components, amounting to approximately 3.5 percent of collections. A total of 184 of 203 positive samples were confirmed by culture and identification of an aliquot from the BacT/ALERT bottle, indicating that 19 cultures were false-positive owing to BacT/ALERT incubator malfunction. Furthermore, because the original product was not retested, false-positive samples due to contamination of the sample at the time of initial testing were not excluded. Bacterial contaminants were predominantly skin organisms, including 9.8 percent (18 of 184) anaerobic bacteria (9 Propionibacterium acnes and 9 gram-positive cocci, not otherwise specified). A total of 113 of 203 (56%) of the components had been transfused by the time of product withdrawal without recorded adverse events, leading the authors to question the value of bacterial testing. They also report two confirmed septic reactions due to Bacillus cereus contamination from pools that had tested culture-negative, indicating a significant false-negative rate (1 in 70,260 individual buffy-coat preparations). Holme and colleagues2 report the results of 118,067 cultures of apheresis and WBD PLTs at 23 regional blood centers utilizing the Pall eBDS system. Three to four milliliters of PLT product were inoculated into a dedicated culture bag 24 hours after collection and incubated for 24 to 30 hours before bacterial growth was assayed with an oxygen sensor. They report 118 of 118,067 (1 in 1,001) initial positive tests with 23 (1 in 5133) results confirmed positive by reculture of the primary product. All true-positive samples were skin organisms (Staphylococcus sp. or Streptococcus sp.). A total of 95 of 118 (80.5%) were false-positive results (1 in 1243 cultures) and a single false-negative (i.e., where the product was cleared for clinical use by eBDS testing but subsequently was shown to be contaminated) was reported in which a Staphylococcus epidermidis contamination was apparently missed. Fang and coworkers1 report the largest series to date of 350,658 cultures of apheresis PLTs performed at 36 American Red Cross regional blood centers utilizing a standardized protocol. Samples of 3 to 5 mL taken from apheresis PLTs at least 24 hours after collection were inoculated under aseptic conditions into BacT/ALERT aerobic bottles and monitored for 5 days. Products were released after 12 hours of incubation. They show a 1 in 1,601 initial positive rate of which 30.6 percent were confirmed positive on retesting of the original product, for an overall confirmed positive rate of 1 in 5234 collections. Unconfirmed positive samples (69.4% of initial positive samples, 1 in 2307 collections) were attributed to BacT/ALERT instrument errors (1 in 9740), to contamination at the time of sampling (1 in 4614), or to an undetermined etiology (indeterminate; 1 in 8766). They also report three septic transfusion reactions, owing to coagulase-negative Staphylococcus, including one fatality, that were not prevented by bacterial testing, suggesting a false-negative rate of 1 in 116,886 collections. Taken together, the Pall eBDS and BacT/ALERT systems rendered practically identical results of approximately 1 in 5000 confirmed positive samples, when false-positive samples (which were twice as frequent [1 in 1243 vs. 1 in 2307] with the eBDS system) were rigorously excluded. These rates are lower than the approximately 1 in 3000 confirmed positive rate predicted by prior smaller studies.6, 7 Clinical septic reactions to tested products were documented in all three reports at rates ranging from 1 in 70,000 to 1 in 118,000.1, 8 This rate is also lower than the approximately 1 in 25,000 (range, 1 in 15,000 to 1 in 100,000) rate predicted by prior studies utilizing untested products.7, 9,10 In one of these studies, Ness and associates,10 with careful monitoring of transfusion recipients, found a constant risk of septic PLT transfusion reactions of 1 in 15,000 donor exposures, regardless of whether single-donor PLTs or WBD PLTs were transfused. Interestingly, their results closely predicted the actual incidence of bacterial contamination if approximately one-third of bacterially contaminated PLT concentrates result in septic reactions. In summary, these data suggest that bacterial culture has reduced but not eliminated bacterial contamination and that there persists a significant and as yet unquantified residual contamination rate after testing with the eBDS and BacT/ALERT systems. With culture-based systems, there are at least three ways to increase analytical sensitivity and to decrease the time to a positive test: delay sampling to allow the proliferation of low-titer bacteria in the product, increase sample volume to decrease sampling error, and culturing under a variety of conditions (e.g., both aerobic and anaerobic). These manipulations would have the consequences of further reducing available shelf life, increasing sampling losses, and increasing price, respectively. The success of bacterial testing has naturally led to a call for extended (7-day) storage of tested products, and the FDA has set two conditions: The products must be stored in containers approved for 7-day storage (these currently exist for both apheresis and WBD PLTs) and PLTs must be tested with an FDA-approved bacterial detection “release test.”11 The latter criterion may not be easy to meet in practice. To consider the current QC tests as release tests, the FDA requires that a standardized method be utilized with respect to sample timing, size, and duration of culture and that a clinical validation study be performed to define the tests’ sensitivity, specificity, and positive and negative predictive value. Working with the AABB bacterial detection task force, the FDA defined the goals for extended storage to be evidence that a sample tested on Day 1 renders the residual risk of bacterial contamination on Storage Day 7 to be less than the current risk of untested PLTs with 5 days of storage (now recognized as approx. 1 in 5000). A sample size of 50,000 tests is estimated to be necessary to provide 95 percent confidence that if a contamination rate of 1 in 10,000 is observed, the actual rate is less than 1 in 5000. Subsequently, the FDA indicated that it would consider allowing validation of a bacterial release test during routine clinical practice in a postmarketing study. As noted, Gambro BCT has now received approval to initiate such a study in the Fall of 2005.3 Apheresis PLTs collected on Gambro Spectra or Trima apheresis platforms will be cultured 24 hours after collection utilizing the BacT/ALERT system aerobic and anaerobic cultures. Products will be released for clinical use after 24 hours of incubation and may be labeled for 7-day storage. All products that are not used clinically will be recultured under the same conditions on Day 8 after collection. The FDA will require 50,000 outdate cultures to validate formally the BacT/ALERT system as a PLT release test, provided that 5 or fewer (1 in 10,000) true-positive Day 8 cultures are detected. Other manufacturers are currently considering approaching the FDA with variations of this protocol for similar validation. We need to consider the unintended consequences of this extended storage study: Compared to the protocol described by Fang and coworkers,1 PASSPORT would incur the additional cost of anaerobic cultures and further delay product release by 12 hours to gain 48 hours of shelf life, a net gain of 36 hours. Driven by the desire to decrease outdates and to utilize current resources better, transfusion services will likely continue to issue older PLTs first. Older PLTs have reduced recovery and survival (quantified by Dumont and colleagues12 as a decline from 63.0% to 53.9% recovery and 161- to 133-hr survival from Day 5 to Day 7, respectively). This in turn may lead to reduced transfusion interval and increased number of transfusions, increasing the overall demand for PLTs. It would also expose patients to a bacterial contamination risk that is as yet unquantified. Blood centers may also view extended storage as an opportunity to reduce weekend collections, a practice that has been driven by the current limited shelf life, which could further reduce inventories. The net effect of these factors is difficult to predict but will almost certainly affect transfusion services differentially. Large services that maintain substantial inventories will see the benefit of cost savings and increased availability from reduced outdates, whereas smaller services that do not maintain inventories may see no benefit, only the adverse consequences of increased cost. The PASSPORT study will define the sensitivity of the BacT/ALERT assay, the false-negative rate, and the residual risk of bacterial contamination of products tested on Day 1 (assuming that the false-negative rate is > 1 in 50,000). Indeed, false-negative samples due to low-level contamination and slow-growing organisms are highly likely to be detected. This is implied by the continuing occurrence of septic reactions with tested products and preliminary reports that suggest that as many as half of contaminated products may be missed by Day 1 cultures.13-15 Of particular concern is P. acnes, a slow-growing anaerobic skin contaminant that grows poorly under the aerobic conditions of PLT storage. P. acnes has rarely been implicated in PLT septic reactions with 5 days of storage but may be more dangerous after 7 days.13 Should the residual contamination rate with this organism be more frequent than the FDA's requirements allow for a bacterial release test (defined as a rate of 1 in 10,000 above), we will be left with the question of whether this is clinically significant and what sacrifices are appropriate to reduce the risk level further, in terms of sampling delay, product loss, and cost. There are alternatives to extended storage, if the goal is to increase PLT availability. Increased apheresis PLT collections in response to increased demand is a simple response that should not be ignored. Switching to point-of-issue bacterial QC tests would allow earlier release of PLTs and therefore a 24-hour increase in hospital shelf life of younger PLTs, even with 5-day storage. No practical point-of-issue tests are yet FDA-licensed for this purpose, and tests in development generally have bacterial detection limits in the 103 per mL range,16 raising the issue of what constitutes an acceptable level of contamination. To our mind, the technologic development that would have the greatest impact on PLT availability would be prestorage pooling of WBD PLTs, an available source of PLTs that is mostly discarded in the United States. Prestorage pooling of PLTs derived from the PLT-rich plasma or buffy-coat production method may be subjected to bacterial testing by the current licensed QC tests. The absence of recruitment and donation costs would render bacterially tested, prestorage pooled WBD PLTs a highly competitive product, and at least one manufacturer (Pall Corp., Covina CA) is exploring this option.17 Finally, it should be mentioned that there are available technologies that may render PLTs free of nonsporulating bacteria, as well as viruses and other pathogens: These include the use of amotosalen and ultraviolet (UV) light (Intercept pathogen reduction system, Cerus Corp., Concord CA) and the use of riboflavin plus UV light (Mirasol, Navigant Biotechnologies, Lakewood, CO). The Intercept system is already in use in Europe.18, 19 Treated products may be released into inventory within 24 to 36 hours of collection, offering a 4-day shelf life even with 5-day storage. Validation studies for extended storage have been accepted for CE mark by the European regulatory authorities (L. Corash, personal communication), promising an effective 6-day shelf-life in the transfusion service. Mirasol is not presently available for licensed use; however, PLT function has been demonstrated to be well preserved for 5 days with this technology.19 The major disadvantage of both systems is that in studies to date there has been an approximately 12 to 30 percent relative loss of PLT recovery.18-21 Although this will require an increased PLT dose for many patients, it will be interesting to see careful cost studies of pathogen-reduced PLTs compared to current products. In particular, pathogen-reduced pooled WBD PLTs may be no more costly than cultured apheresis concentrates. In conclusion, bacterial testing of apheresis PLTs stored for 5 days has undoubtedly reduced the risk of bacterial contamination and increased patient safety. There is an immediate need for extended storage to help hospitals manage their inventories and reduce outdates. It is ironic that this will come at a price of increased potential bacterial sepsis risk and the use of less efficacious, older products. The prospect of broad pathogen reduction offers a promising alternative to the current situation, given a diminishing tolerance for the risk of residual bacterial contamination.