DONOR INFECTIOUS DISEASE TESTING Evaluation of a rapid colorimetric assay for detection of bacterial contamination in apheresis and pooled random-donor platelet units W. Andrew Heaton,1 Caryn E. Good,2,3 Rakijah Galloway-Haskins,1 Roslyn A. Yomtovian,2,4 and Michael R. Jacobs2,3
BACKGROUND: Despite existing strategies, bacterial contamination of platelets (PLTs) remains a problem, and reliable testing near the time of use is needed. We evaluated the BacTx assay (Immunetics, Inc.), a rapid colorimetric assay for detection of bacterial peptidoglycan, for this purpose. STUDY DESIGN AND METHODS: Apheresis- and whole blood–derived PLT units, the latter tested in 6-unit pools, inoculated with 10 representative bacterial species (eight aerobic, two anaerobic), were tested with the BacTx assay at two sites to determine analytic sensitivity and time to detection. Specificity on sterile PLTs and reproducibility across different PLT units and assay kit lots was also determined. RESULTS: Analytical sensitivity for the 10 bacterial species ranged from 6.3 × 102 to 7.6 × 104 colonyforming units (CFUs)/mL. In time-to-detection studies after inoculation of PLTs with 0.7 to 5.3 CFUs/mL, 10 replicates of all eight aerobic species were positive when bacterial titers were above the analytic sensitivity detection limit, which occurred at 48 hours for 60 PLT units and at 72 hours for the remaining 4 units, as well as at 7 days for all units. Specificity was 99.8% and reproducibility was 100%. CONCLUSIONS: The BacTx assay had an analytical sensitivity below the 105 CFUs/mL threshold of clinical significance, detected all eight aerobic bacterial species 48 to 72 hours after inoculation as well as at 7 days, and had high specificity and reproducibility. These findings suggest that the BacTx assay will be a valuable test for detection of clinically relevant levels of bacterial contaminants in PLT units and pools near time of use.
1634
TRANSFUSION Volume 54, June 2014
B
acterial contamination of platelets (PLTs) continues to be the most significant infectious cause of transmission-associated morbidity and mortality, with an estimated frequency of 1 in 2000 to 3000 units, compared to rates for hepatitis B, hepatitis C, and human immunodeficiency viruses of 1 in 350,000, 1,149,000, and 1,467,000 transfusions, respectively.1-6 Detection is complicated by the low number of bacteria that contaminate the donation during the collection process and the variable growth of bacteria in PLTs at room temperature storage conditions.7 In spite of improved arm preparation methods and diversion of the initial donation, both of which have decreased the incidence of bacterial contamination, a significant residual risk remains and methods to limit and detect this have been required by the AABB (Advancing Transfusion and Cellular Therapies Worldwide) since 2004 (AABB Standard 5.1.5.1, effective March 2004).8 This has been
ABBREVIATIONS: BacTx assay = peptidoglycan assay using prophenoloxidase cascade system; LRAP = leukoreduced, apheresis platelet; LRWBDP = leukoreduced, whole blood–derived platelet. From the 1Blood Component Research Laboratory, Feinstein Institute for Medical Research, Manhasset, New York; and the 2 Department of Pathology, Case Western Reserve University; the 3 Department of Pathology, University Hospitals Case Medical Center; and the 4Department of Pathology and Laboratory Medicine, Louis Stokes VA Medical Center, Cleveland, Ohio. Address reprint requests to: Michael R. Jacobs, MD, PhD, Department of Pathology, University Hospitals Case Medical Center, 11100 Euclid Avenue, Cleveland, OH 44106; e-mail: [email protected]. This study was supported in part by Immunetics, Inc. Received for publication August 14, 2013; revision received October 4, 2013, and accepted October 7, 2013. doi: 10.1111/trf.12603 © 2014 AABB TRANSFUSION 2014;54:1634-1641.
BACTX ASSAY FOR BACTERIA IN PLTs
accomplished by culture using broth cultures of a sample of apheresis PLTs and, where pooled at blood centers, pooled whole blood–derived PLTs taken approximately 24 hours postdonation, associated with sample reading 36 to 48 hours postdonation and release of units to transfusion centers if culture results are negative at this time.9 Testing of whole blood–derived PLT concentrates pooled immediately before release, which was previously allowed using simple but insensitive and nonspecific pH or glucose measurements, is now required to be performed using specific Food and Drug Administration (FDA)-approved bacterial assays (AABB Standard 5.1.5.1.1, effective January 2011).10 These measures have helped lower the number of fatalities due to bacterial contamination of PLT products reported to the FDA from approximately seven per year to two to three per year, predominantly associated with reductions in deaths due to Gram-negative bacterial species.11 Detection of bacterial contamination of PLT products by bacterial culture as currently performed is limited by the low bacterial load present at the time of collection.12 Initial studies of bacterial contamination were modeled on bacterial loads of 0.154 colony-forming units (CFUs)/ mL, or approximately 50 CFUs per apheresis product, where detection by culture of 4- to 8-mL aliquots was likely. However, later studies suggested that loads 1/10 of this were not uncommon in many units.2,7,13,14 Since industry standards include the sampling of approximately 8 mL of PLT concentrates or approximately 2% of the product volume, it has been estimated by Poisson sampling analysis that current culture techniques would detect less than 50% of contaminated units.15,16 Other limitations of early culture include the variable rates of bacterial growth at low bacterial loads and the fact that only anaerobic culture (not now a blood center standard) detected low levels of common skin contaminant Gram-positive cocci.15,16 All these limitations of the current culture approach toward minimizing the hazard of bacterial contamination of PLT concentrates are reflected in a recent study suggesting that less than 50% of bacterially contaminated apheresis units are detected by early culture.2 For all these reasons a sensitive method for testing apheresis and whole blood–derived PLT units at time of issue is needed. Two methods are now FDA-approved and meet regulatory requirements—the PGD test (Verax Biomedical, Marlborough, MA), and the BacTx assay (Immunetics, Inc., Boston, MA), which are both cleared for quality control testing of apheresis units and pools of up to 6 whole blood–derived units.17,18 The Immunetics BacTx assay, intended for use as a qualitative test for bacterial contamination of PLT products, detects peptidoglycan, a universal component of cell walls of both Grampositive and Gram-negative bacteria.19 The assay is based on an insect innate immune system bacterial defense mechanism, in which peptidoglycan-binding proteins
from insect hemolymph trigger a prophenoloxidase serine protease cascade. The phenoloxidase enzyme activity converts catechols to quinones, which are complexed with MBTH, a soluble chromogen that turns red upon quinone binding, generating measurable absorbance at 500 nm. Absorbance is monitored by a dedicated photometer over a fixed time period. This study examined the analytical sensitivity and time to detection of the BacTx assay using PLT units inoculated with 10 representative bacterial species. Specificity and reproducibility across test kit lots were also studied. All testing was done at two study sites, Cleveland, Ohio, and Manhasset, New York.
MATERIALS AND METHODS PLT units All testing was performed at two study sites using indate, leukoreduced, apheresis PLT (LRAP) units and leukoreduced, whole blood–derived PLT (LRWBDP) units acquired from blood centers. LRWBDP minipools were prepared as needed by pooling 2-mL volumes from LRWBDP units, using sterile transfer methods, with concurrent culture screening to verify sterility of units.
Bacterial strains Bacteria used in the analytical sensitivity, time to detection, and reproducibility assays included the following 10 bacterial species (ATCC number): aerobic or facultative Gram-positive species, Staphylococcus aureus (27217), Staphylococcus epidermidis (49134), Bacillus cereus (11778), and Streptococcus agalactiae (12386); aerobic or facultative Gram-negative species, Serratia marcescens (43862), Pseudomonas aeruginosa (27853), Escherichia coli (25922), and Klebsiella oxytoca (43863); and anaerobic Gram-positive species, Clostridium perfringens (3629) and Propionibacterium acnes (11827).
BacTx assay The BacTx assay (Immunetics, Inc.) for bacterial peptidoglycan is provided as a kit comprising all necessary reagents to carry out multiple tests. A photometer is provided as a required accessory and has positions for up to eight tests to be run simultaneously. The photometer is attached to a dedicated personal computer that records information on specimens being tested and results obtained from the photometer. Other equipment required includes a microcentrifuge capable of 14,000 to 20,000 × g, a variable-speed mixer, and micropipettors capable of delivering up to 1000-μL volumes. To perform the BacTx assay with LRWBDP, sample volumes of 0.5 mL are sterilely obtained from PLT pools to be tested and placed in Volume 54, June 2014
TRANSFUSION
1635
HEATON ET AL.
2-mL microcentrifuge tubes containing 1.0 mL of lysis buffer. Microcentrifuge tubes are then immediately capped, inverted three times to mix, and centrifuged at 18,000 × g for 3 minutes. For LRAP, 1.0 mL of PLT are combined with 0.75 mL of lysis buffer and mixed vigorously before centrifugation. For all PLT types, the supernatant is then decanted into a waste container and any remaining droplets are removed by tapping tubes against the rim of the waste container. The remaining pellets are then resuspended in 0.5 mL of extraction reagent by pipetting up and down 10 times. Volumes of 0.5 mL of neutralization reagent are placed in 1.5-mL microcentrifuge tubes and the resuspended pellets are transferred to these tubes, which are then capped and inverted three times to mix. Volumes of 0.3 mL are then transferred to BacTx reaction tubes, which are then briefly mixed vigorously and placed in the BacTx reader. The BacTx reader cycle is started from the attached computer and reactions in the BacTx reaction tubes are monitored by the BacTx reader software for changes in color (absorbance) over 30 minutes. An increase in absorbance within this time period above a predetermined cutoff is interpreted by the software as a positive or “fail” result, indicating bacterial contamination. Otherwise, a negative or “pass” result is delivered at the end of the time period. Due to nonlinear signal amplification the signal is not directly related to input bacterial concentration, but the time to exceed the assay cutoff is affected by bacterial input concentration.19 Results are stored electronically in an assay log file, which can also be printed and exported to a laboratory information system. The turnaround time for testing a batch of up to eight PLT samples is less than 45 minutes.
Analytical sensitivity The analytical sensitivity of the BacTx assay was determined in LRAP and freshly prepared 6-unit LRWBDP minipools in tubes using bacterial suspensions prepared at targeted counts. Bacterial species were grown overnight on aerobic or anaerobic blood agar plates and then inoculated into appropriate broth (tryptic soy broth or brain heart infusion broth for aerobes; anaerobic Brucella broth for anaerobes) and incubated at 35°C for several hours until target optical density (OD) values were obtained. Bacterial suspensions were then diluted in phosphatebuffered saline (PBS) to defined bacterial load levels calibrated by OD against a standard curve to allow initial dilution in LRAP and LRWBDP pools (pools prepared from 2-mL aliquots from 1 inoculated and 5 uninoculated units to form 6-unit minipools of 12 mL) at defined bacterial titers. These targeted bacterial titers in LRAP and LRWBDP pools were 1 × 103, 5 × 103, 1 × 104, 2 × 104, 4 × 104, and 8 × 104 CFUs/mL. Ten replicates of LRAP and LRWBDP minipools containing each species and concentration 1636
TRANSFUSION Volume 54, June 2014
were tested by BacTx assay and interpreted as detected at the lowest bacterial titer at which all 10 replicates tested positive. Quantitative bacterial culture was performed on samples from each PLT or pool by plating 0.1 mL of serial 10-fold dilutions in duplicate onto blood agar plates to determine actual bacterial titers in the PLT or pool at the time of testing.
Time to detection This was evaluated by inoculating very low titers (target, 1-5 CFUs/mL) of each of the 10 bacterial species into LRAP and individual LRWBDP units. Bacterial species were grown overnight on aerobic or anaerobic blood agar plates and inoculated into appropriate broth and incubated at 35°C to specific levels defined by OD as described above. After dilution in PBS to defined bacterial load levels, 4 LRAP and 4 LRWBDP units were inoculated (two with 0.1 mL and two with 0.3 mL) and sampled by plate culture incubated overnight to determine actual bacterial inocula. An additional unit was inoculated with saline to serve as a negative control. Inoculated units were incubated at 22°C in a PLT incubator-rocker for 7 days. On Day 1, (t = 24 hr postinoculation), the bacterial inocula were determined from growth plates, and two of each set of 4 inoculated units were chosen for further study based on the bacterial inocula being within the targeted range of 1 to 5 CFUs/mL. Aliquots were removed from the inoculated units selected for study and 0.1 mL plated onto blood agar plates. If growth was detected on these plates after incubation for 24 hours (t = 48 hr), indicating that bacterial growth had occurred, one of the inoculated units was selected for further study and was incubated for a total of 7 days (168 hr). On Day 2 (t = 48 hr) and Day 7 (t = 168 hr), an aliquot from inoculated units in the case of LRAP units, or from a LRWBDP pool prepared from an inoculated LRWBDP unit pooled with aliquots from 5 fresh sterile LRWBDP units, was tested by BacTx assay in replicates of 10. In addition the saline-inoculated unit was tested by BacTx assay in triplicate at each time point. If fewer than 10 replicates from a unit were positive at t = 48 hours the unit was retested at t = 72 hours. Quantitative bacterial counts were done by plate culture at t = 0 on inocula and on the inoculated units or pools on each day of BacTx assay testing. The bacterial species used were identified independently from the initial inoculum and after recovery on Day 7 to assure that the species detected were those inoculated.
Specificity Specificity testing was performed on uninoculated LRAP and uninoculated 6-unit LRWBDP minipools. To confirm sterility LRWBDP pools and LRAP units were cultured by plate culture in triplicate on aerobic blood agar plates,
BACTX ASSAY FOR BACTERIA IN PLTs
TABLE 1. Analytical sensitivity of the BacTx assay in LRAP units and LRWBDP minipools containing bacteria added at titers ranging from 1 × 103 to 8 × 104 CFUs/mL, with actual bacterial titers determined by quantitative plate culture at time of testing*
Bacterial species B. cereus C. perfringens E. coli K. oxytoca P. acnes P. aeruginosa S. marcescens S. aureus S. epidermidis S. agalactiae
Lowest bacterial titer (CFUs/mL) at which all 10 samples were positive by BacTx assay by PLT type and test site LRAP LRWBDP Site 1 Site 2 Site 1 Site 2 1.9 × 103 1.3 × 103 1.7 × 103 1.4 × 103 9.4 × 102 4.8 × 103 2.8 × 103 4.5 × 103 7.6 × 104 3.3 × 104 5.1 × 103 8.7 × 103 1.6 × 104 7.3 × 103 6.8 × 103 9.9 × 103 5.0 × 103 8.5 × 103 7.2 × 103 1.1 × 103 2.7 × 104 2.0 × 104 9.6 × 103 5.0 × 104 4.2 × 103 5.3 × 103 5.8 × 104 6.7 × 103 2.2 × 103 1.1 × 103 2.1 × 103 4.0 × 103 1.3 × 103 6.3 × 102 2.0 × 103 2.4 × 103 4.5 × 103 3.3 × 103 3.6 × 103 2.7 × 104
* LRWBDP minipools were prepared from one LRWBDP unit containing bacteria pooled with 5 uninoculated LRWBDP units. Ten replicates of the BacTx assay were performed for each bacterial species and concentration and analytic sensitivity defined as the lowest bacterial titer at which all 10 samples were positive by BacTx assay.
using 0.1-mL volumes. LRAP units were also cultured by an automated bacterial detection system (BacT/ALERT, bioMérieux, Durham, NC) in BPA and BPN bottles, using 4-mL aliquots per bottle. For LRWBDP, on each day of testing at each site, 8 individual PLT units were used to prepare 24 unique minipools containing 1-mL aliquots from 6 different units, organized so that each minipool had a set of six unique constituents. A different set of 8 PLT units was used for each of the nine testing iterations at each site, with a total of 144 LRWBDP units used and 432 unique minipools (216 at each site) prepared. On each day of testing the 24 unique minipools were tested in three groups of eight using three different BacTx kit lots at the two testing centers (three runs of 24 pools per lot per site using one lot per day). LRAP units were tested with five BacTx kit lots. For LRAP, a retest algorithm was used to determine assay reactivity. LRAP units that were initially reactive in the BacTx assay were retested in duplicate. If either retest was positive, the test result was confirmed as reactive. If neither retest was reactive, the unit was designated as nonreactive.
Reproducibility Interlot and intersite reproducibility were tested using frozen aliquots of 10 bacterial suspensions in saline of the 10 bacterial species used above and a saline negative control. The aliquots contained bacteria at titers that were 0.5 to 1.5 log CFUs/mL higher than the sensitivity determined for a given strain in the analytical sensitivity portion of the study and were frozen in tubes in 100-μL volumes. For LRAP, a test panel was used for testing at the two sites using three lots of BacTx assay kits on 3 different days. The test panel consisted of 10 bacterial members and one negative member. The sterility of LRAP used in
the study was confirmed by either BacT/ALERT culture or agar plate culture. Similarly, for LRWBDP, three different BacTx kit lots were used at both sites on 3 different days. On each testing day, four unique 18-mL PLT pools were prepared using a combination of 3-mL aliquots from 6 of 7 individual LRWBDP units. Aliquots of 0.1 mL from each pool were placed on blood agar plates and incubated aerobically to assure sterility of the PLT pools. Aliquots of 0.5 mL from each LRWBDP pool or 1.0 mL from each LRAP unit were added to each of the 10 thawed bacterial suspension aliquots and a saline control, and then tested by BacTx assay.
RESULTS Analytical sensitivity The analytic sensitivity of the BacTx assay on LRAP and 6-unit LRWBDP minipools inoculated with the 10 representative PLT bacterial contaminants studied ranged from 6.3 × 102 CFUs/mL for S. epidermidis to 7.6 × 104 CFUs/ mL for E. coli (Table 1). Overall, five of the bacterial species were detected at less than 104 CFUs/mL in both PLT types at both sites, while the remaining species were detected at not more than 7.6 × 104 CFUs/mL.
Time to detection LRAP units were inoculated with 1.3 to 5.3 CFUs/mL and LRWBDP units with 0.7 to 5.0 CFUs/mL. Time to detection in LRAP and 6-unit LRWBDP minipools, prepared at each time of testing from 1 inoculated and 5 sterile units, is shown in Table 2. Bacterial detection in all 10 replicates was found at 48 hours for 28 of the 32 challenges with the eight aerobic or facultative challenge bacterial species, with detection in all 10 replicates in the remaining four Volume 54, June 2014
TRANSFUSION
1637
HEATON ET AL.
TABLE 2. Results of time to detection studies after inoculation of LRAP and LRWBDP units with aerobic bacterial species* Number of replicates (out of 10) detected by BacTx assay by PLT type and time point LRWBDP Bacterial species E. coli P. aeruginosa K. oxytoca S. marcescens B. cereus S. aureus S. epidermidis S. agalactiae
Site number 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2
48 hr 10 10 10 10 10 10 10 10 10 10 10 10 9 10 10 10
LRAP 72 hr —† — — — — — — — — — — — 10 — — —
48 hr 10 10 0 0 10 10 10 10 10 10 10 10 10 10 10 0
72 hr — — 10 10 — — — — — — — — — — — 10
* LRAP units were inoculated with 1.3 to 5.3 CFUs/mL bacteria at Time 0. LRWBDP units were inoculated with 0.7 to 5.0 CFUs/mL of bacteria at Time 0, with inoculated units pooled with 5 uninoculated LRWBDP units at time of testing. Ten replicates were tested by BacTx assay at 48 hours. If one or more of the 10 aliquots was negative at 48 hours, testing was repeated at 72 hours. † Not tested at 72 hours as all 10 replicates were positive at 48 hours.
challenges at 72 hours. The challenge organisms in these four cases where detection occurred at 72 hours were P. aeruginosa in LRAP units at both sites; S. epidermidis in a LRWBDP minipool at one site, where bacterial contamination was detected by the BacTx assay in nine of 10 replicates at 48 hours at a bacterial titer of 6.6 × 102 CFUs/mL; and S. agalactiae in LRAP at one site (Table 2). Bacterial titers in pools at time of detection of all 10 replicates ranged from 8.7 × 102 CFUs/mL for S. agalactiae to 6.4 × 108 CFUs/mL for E. coli. The two challenge anaerobic species studied, P. acnes and C. perfringens, failed to grow in inoculated PLT units and remained undetectable at all time points tested by both plate culture and BacTx assay.
TABLE 3. Specificity of the BacTx assay in sterile LRAP and LRWBDP at the two test sites BacTx assay result Negative Initially reactive Repeat reactive
LRAP units (n = 505) 501 4 1
LRWBDP pools (n = 432)* 431 1 ND
* Prepared from 144 LRWBDP units. ND = not determined.
tests performed had to be retested due to an aborted assay run.
Reproducibility Specificity A total of 432 unique LRWBDP pools, derived from 144 LRWBDP units, were tested (216 at each site) in three groups of 8 on 9 separate days using three different BacTx kit lots. There was no bacterial growth associated with any of these pools. Of these 432 pools, 431 tested correctly as negative, for a test specificity of 99.8% (the one false positive was not retested) (Table 3). For LRAP, a total of 505 units were tested (96 units at Site 1 and 409 at Site 2) with five BacTx kit lots. Of the 505 units tested, 501 were negative on initial testing, and four were initially reactive. Upon retesting, only one of the seven initially reactive samples was reactive (Table 3). Therefore, 504 of 505 units were initially negative or nonreactive on repeat testing for a test specificity of 99.8% for LRAP. One of the 937 specificity 1638
TRANSFUSION Volume 54, June 2014
A total of 528 reproducibility tests were run on LRAP and LRWBDP pools (132 assays with each PLT type at each site), using three different BacTx kit lots per PLT type (six kit lots in total) on 3 different days on the 11 PLT specimens prepared. In all 480 cases where bacterial suspensions had been added, the BacTx assays on the PLT specimens yielded positive results, while the 24 LRAP and 24 LRWBDP pool specimens inoculated with saline yielded negative results (Table 4). In total, all 528 BacTx assays yielded correct results, for 100% reproducibility.
DISCUSSION There is considerable interest in addressing the continuing problem of bacterial contamination of PLT products.20
BACTX ASSAY FOR BACTERIA IN PLTs
TABLE 4. Reproducibility of the BacTx assay using three BacTx assay kit lots at the two test sites Lot 1 2 3
LRAP 88/88* 88/88 88/88
LRWBDP 88/88 88/88 88/88
* Results are expressed as number of tests with expected result/number of tests performed.
A 1-day public conference on this topic was sponsored by AABB in July 2012 with the objectives of discussing the residual risk of bacterial contamination in apheresis PLTs, describing the use of secondary bacterial testing of PLT components using point-of-issue tests, and hearing practical experiences from transfusion services using these tests with whole blood–derived PLTs.21 Presenters at this meeting noted that, although there is no FDA requirement for testing of PLT products for bacterial contamination, several products have been approved by the FDA to decrease the risk of bacterial contamination, including diversion pouches and culture systems such as BacT/ ALERT and eBDS, as well as point-of-issue tests such as the PLT PGD test and the BacTx assay system. On September 21, 2012, the Blood Products Advisory Committee of the FDA met to discuss considerations for strategies to further reduce the risk of bacterial contamination in PLTs.22,23 The committee voted unanimously in favor of additional measures being necessary to decrease the current risk of transfusion of bacterially contaminated PLT products. The committee voted unanimously against reducing PLT product shelf life from 5 to 4 days as being insufficient to decrease the risk of transfusion-associated septic reactions and to obviate the need for additional testing. The committee also voted in favor of a strategy of culturing PLTs after the first 24 hours of storage, as currently practiced, and then retesting on Day 4 or Day 5 just once with a rapid test on the day of transfusion; however, the vote was split on applying this strategy to retesting on Day 3. The FDA is now considering issuing guidance on additional testing requirements for PLTs near time of issue. The BacTx assay is a simple rapid test for bacterial contamination in a tube format that is suitable for near time of issue use and can be performed in under 1 hour. In our studies the assay successfully detected 10 representative bacterial species inoculated into representative PLT concentrates or pools, with detection at concentrations ranging from 6.2 × 102 to 7.6 × 104 CFUs/mL in LRAP and pooled LRWBDP. The threshold of bacterial detection of the assay was therefore below the bacterial load of more than 105 CFUs/mL reported to be associated with septic transfusion reactions.24 Compared with the analytic sensitivity values for a similar range of species of approximately
104 to 106 CFUs/mL reported for the PGD test, the BacTx assay was more sensitive overall, by more than an order of magnitude in several instances.18,25 For example, in LRAP, the BacTx assay was more than 100-fold more sensitive than the PGD test for detection of S. marcescens (5.3 × 103 CFUs/mL vs. 8.2 × 105 CFUs/mL) and more than 10-fold more sensitive for detection of S. agalactiae (4.5 × 103 CFUs/mL vs. 5.5 × 104 CFUs/mL) and C. perfringens (4.8 × 103 CFUs/mL vs. 8.9 × 104 CFUs/ mL).18,25 Of the 10 bacterial strains used in the current work, only one, S. marcescens ATCC 43862, was the same strain of this species used in evaluations of analytic sensitivity with the PGD test. Comparisons between the analytic sensitivity values found in this and the PGD test studies are therefore confounded by the use of predominantly different strains that may have distinct growth characteristics, making direct comparisons less precise. The bacterial species studied included four each of aerobic or facultative Gram-positive and Gram-negative species, representing the more significant organisms reported to be associated with posttransfusion bacterial reactions.11,20,26 In addition two anaerobic Gram-positive species also associated with transfusion reactions were detected, although the anaerobic bacteria failed to grow in the PLTs, so time to detection could not be evaluated. In the PLT units spiked with inocula between 0.7 and 5.3 CFUs/mL, 28 of the 32 challenges with the eight aerobic species were detected at 48 hours, with the remaining four challenges detected at 72 hours. These findings are similar to those found in time to detection studies using the PGD test,18 in which three of the strains studied, S. marcescens ATCC 43862, S. aureus ATCC 27217, and S. epidermidis ATCC 49134 were the same as those used in our time to detection evaluations of the BacTx assay. Publications based on the implications of time to culture positivity have suggested that, while the median number of bacteria initially contaminating PLT units may be as high as 50 CFUs/product, very low loads of approximately 5 CFUs/product are more common.2,14,15,27 Additionally, bacteria may not start growing immediately and can have an extended lag phase. Under such circumstances the time testing is performed, the length of time needed to generate a result and the limit of detection of the test method are important variables. Although recent reports based on bacterial growth curves with culture using time to positive modeling have suggested that inoculation of larger samples into culture bottles might increase the probability of detection of contamination, these studies do not address the problems posed by delayed bacterial growth in PLT with low inocula. The relatively low assay sensitivity threshold and short time to detection of contamination reported for the BacTx assay in the current study suggest that application of this assay as a point-of-release test might confer an additional level of transfusion safety.2,14,15,28 Volume 54, June 2014
TRANSFUSION
1639
HEATON ET AL.
Specificity with multiple lots and technical operators was high, at more than 99.8%, with a very low invalid run rate (one of 937 tests). Reproducibility across different lots and with different operators was likewise excellent at 100%, and equipment failures were minimal. As such the BacTx assay appears well suited for use in a routine laboratory environment. In this evaluation, the BacTx assay demonstrated the highest levels of sensitivity and specificity reported for a cleared commercial rapid test for detection of bacterial contamination in PLTs. The assay would thus serve the need for point-of-release bacterial screening of apheresis and pooled random-donor PLT concentrates at 3 or more days of room temperature storage to prevent transfusion of PLT units and pools contaminated with clinically significant levels of bacteria.3,27 The use of this test to extend the shelf life of PLTs beyond the current 5-day period should also be explored. Clinical trials are warranted to determine the value of this product in detecting bacterial contamination of PLT units administered to patients. ACKNOWLEDGMENTS The authors are grateful for technical support from Munira Mulani, MT. The authors acknowledge the assistance of the Immunetics Support Team, Andrew Han, PhD, Neil Krueger, PhD, and Richard Pinkowitz.
CONFLICT OF INTEREST All authors except MRJ, RAY, and WAH have declared no other conflict of interest or financial involvement with this article. MRJ has received research support and/or honoraria from Verax, Pall, Gambro, Hemosystem, Immunetics, Genprime, and Charles River Labs and has been a consultant for BioSense Technologies and Lynntech, Inc. RAY has received honoraria and/or consulting fees from Immunetics, Verax, Pall, Gambro, Hemosystem, Fenwal, and Genprime. WAH has received research support/honoraria support from Verax, Immunetics, Fenwal, Beckman Coulter, Novartis Diagnostics, TerumoBCT, and Haemeonetics.
REFERENCES 1. Dodd R, Kurt Roth W, Ashford P, et al. Transfusion medicine and safety. Biologicals 2009;37:62-70. 2. Eder AF, Kennedy JM, Dy BA, et al. Limiting and detecting bacterial contamination of apheresis platelets: inlet-line diversion and increased culture volume improve component safety. Transfusion 2009;49:1554-63. 3. Jacobs MR, Smith D, Heaton WA, et al. Detection of bacterial contamination in prestorage culture-negative apheresis platelets on day of issue with the Pan Genera Detection test. Transfusion 2011;51:2573-82. 4. Zou S, Dorsey KA, Notari EP, et al. Prevalence, incidence, and residual risk of human immunodeficiency virus and 1640
TRANSFUSION Volume 54, June 2014
hepatitis C virus infections among United States blood donors since the introduction of nucleic acid testing. Transfusion 2010;50:1495-504. 5. Zou S, Musavi F, Notari EP, et al. Prevalence, incidence, and residual risk of major blood-borne infections among apheresis collections to the American Red Cross Blood Services, 2004 through 2008. Transfusion 2010;50: 1487-94. 6. Zou S, Stramer SL, Notari EP, et al. Current incidence and residual risk of hepatitis B infection among blood donors in the United States. Transfusion 2009;49:1609-20. 7. Benjamin RJ, Wagner SJ. The residual risk of sepsis: modeling the effect of concentration on bacterial detection in two-bottle culture systems and an estimation of falsenegative culture rates. Transfusion 2007;47:1381-9. 8. American Association of Blood Banks. Standards for blood banks and transfusion services. 22nd ed. Standard 5.1.5.1. Association Bulletin 03-12, appendix I. 2003. 9. Brecher ME, Jacobs MR, Katz LM, et al. Survey of methods used to detect bacterial contamination of platelet products in the United States in 2011. Transfusion 2013;53:911-8. 10. American Association of Blood Banks. Standards for blood banks and transfusion services. 27th ed. Summary of significant changes. 2011 [cited 2013 Oct 4]. Available from: http://www.aabb.org/sa/standards/Documents/ sigchngstds27.pdf 11. FDA. Fatalities reported to FDA following blood collection and transfusion: annual summary for fiscal year 2012. 2013. [cited 2013 Oct 4]. Available from: http://www.fda .gov/BiologicsBloodVaccines/SafetyAvailability/ ReportaProblem/TransfusionDonationFatalities/ ucm346639.htm 12. Eder AF, Kennedy JM, Dy BA, et al. Bacterial screening of apheresis platelets and the residual risk of septic transfusion reactions: the American Red Cross experience (20042006). Transfusion 2007;47:1134-42. 13. Murphy WG, Foley M, Doherty C, et al. Screening platelet concentrates for bacterial contamination: low numbers of bacteria and slow growth in contaminated units mandate an alternative approach to product safety. Vox Sang 2008; 95:13-9. 14. Pearce S, Rowe GP, Field SP. Screening of platelets for bacterial contamination at the Welsh Blood Service. Transfus Med 2011;21:25-32. 15. Benjamin RJ, Dy B, Perez J, et al. Bacterial culture of apheresis platelets: a mathematical model of the residual rate of contamination based on unconfirmed positive results. Vox Sang 2014;106:23-30. 16. Tomasulo PA, Wagner SJ. Predicting improvement in detection of bacteria in apheresis platelets by maintaining constant component sampling proportion. Transfusion 2013; 53:835-42. 17. Immunetics. BacTx bacterial detection system. 2013. [cited 2013 Oct 4]. Available from: http://www.immunetics.com/ bactx.htm
BACTX ASSAY FOR BACTERIA IN PLTs
18. Verax Biomedical. Platelet PGD Test. 2013. [cited 2013 Oct 4]. Available from: http://www.veraxbiomedical.com/ products/platelet-pgd-test.asp 19. Kovalenko V, Levin A, Beausang L. Rapid peptidogycan-
23. Epstein JS, Vostal JG. FDA contributions to reduction of bacterial contamination in platelet products within the United States. Transfusion 2013;53:232-3. 24. Jacobs MR, Good CE, Lazarus HM, et al. Relationship
based assay for detection of bacterial contamination of
between bacterial load, species virulence, and transfusion
platelets. US Patent 7,598,054. 2009. [cited 2013 Oct 4]. Available from: http://www.google.com/patents/
reaction with transfusion of bacterially contaminated platelets. Clin Infect Dis 2008;46:1214-20.
US7598054 20. Brecher ME, Blajchman MA, Yomtovian R, et al. Addressing the risk of bacterial contamination of platelets within the United States: a history to help illuminate the future. Transfusion 2013;53:221-31. 21. American Association of Blood Banks. Public Conference— Secondary Bacterial Screening of Platelet Components, 7/17/12, Bethesda, MD. 2012. [cited 2013 Oct 4]. Available from: http://www.aabb.org/events/misc/Pages/public -conference.aspx 22. FDA. Meeting of the Blood Products Advisory Committee to discuss considerations for strategies to further reduce the risk of bacterial contamination in platelets, Rockville, MD. 2012. [cited 2013 Oct 4]. Available from: http://
25. Vollmer T, Hinse D, Schottstedt V, et al. Inter-laboratory comparison of different rapid methods for the detection of bacterial contamination in platelet concentrates. Vox Sang 2012;103:1-9. 26. Brecher ME, Means N, Jere CS, et al. Evaluation of an automated culture system for detecting bacterial contamination of platelets: an analysis with 15 contaminating organisms. Transfusion 2001;41:477-82. 27. Kuehnert MJ, Roth VR, Haley NR, et al. Transfusiontransmitted bacterial infection in the United States, 1998 through 2000. Transfusion 2001;41:1493-9. 28. Murphy WG. Pathogen reduction: state of reflection in Ireland. Transfus Clin Biol 2011;18:488-90.
www.fda.gov/AdvisoryCommittees/Calendar/ ucm313863.htm
Volume 54, June 2014
TRANSFUSION
1641
BACKGROUND: Despite existing strategies, bacterial contamination of platelets (PLTs) remains a problem, and reliable testing near the time of use is needed. We evaluated the BacTx assay (Immunetics, Inc.), a rapid colorimetric assay for detection of bacterial peptidoglycan, for this purpose. STUDY DESIGN AND METHODS: Apheresis- and whole blood–derived PLT units, the latter tested in 6-unit pools, inoculated with 10 representative bacterial species (eight aerobic, two anaerobic), were tested with the BacTx assay at two sites to determine analytic sensitivity and time to detection. Specificity on sterile PLTs and reproducibility across different PLT units and assay kit lots was also determined. RESULTS: Analytical sensitivity for the 10 bacterial species ranged from 6.3 × 102 to 7.6 × 104 colonyforming units (CFUs)/mL. In time-to-detection studies after inoculation of PLTs with 0.7 to 5.3 CFUs/mL, 10 replicates of all eight aerobic species were positive when bacterial titers were above the analytic sensitivity detection limit, which occurred at 48 hours for 60 PLT units and at 72 hours for the remaining 4 units, as well as at 7 days for all units. Specificity was 99.8% and reproducibility was 100%. CONCLUSIONS: The BacTx assay had an analytical sensitivity below the 105 CFUs/mL threshold of clinical significance, detected all eight aerobic bacterial species 48 to 72 hours after inoculation as well as at 7 days, and had high specificity and reproducibility. These findings suggest that the BacTx assay will be a valuable test for detection of clinically relevant levels of bacterial contaminants in PLT units and pools near time of use.
1634
TRANSFUSION Volume 54, June 2014
B
acterial contamination of platelets (PLTs) continues to be the most significant infectious cause of transmission-associated morbidity and mortality, with an estimated frequency of 1 in 2000 to 3000 units, compared to rates for hepatitis B, hepatitis C, and human immunodeficiency viruses of 1 in 350,000, 1,149,000, and 1,467,000 transfusions, respectively.1-6 Detection is complicated by the low number of bacteria that contaminate the donation during the collection process and the variable growth of bacteria in PLTs at room temperature storage conditions.7 In spite of improved arm preparation methods and diversion of the initial donation, both of which have decreased the incidence of bacterial contamination, a significant residual risk remains and methods to limit and detect this have been required by the AABB (Advancing Transfusion and Cellular Therapies Worldwide) since 2004 (AABB Standard 5.1.5.1, effective March 2004).8 This has been
ABBREVIATIONS: BacTx assay = peptidoglycan assay using prophenoloxidase cascade system; LRAP = leukoreduced, apheresis platelet; LRWBDP = leukoreduced, whole blood–derived platelet. From the 1Blood Component Research Laboratory, Feinstein Institute for Medical Research, Manhasset, New York; and the 2 Department of Pathology, Case Western Reserve University; the 3 Department of Pathology, University Hospitals Case Medical Center; and the 4Department of Pathology and Laboratory Medicine, Louis Stokes VA Medical Center, Cleveland, Ohio. Address reprint requests to: Michael R. Jacobs, MD, PhD, Department of Pathology, University Hospitals Case Medical Center, 11100 Euclid Avenue, Cleveland, OH 44106; e-mail: [email protected]. This study was supported in part by Immunetics, Inc. Received for publication August 14, 2013; revision received October 4, 2013, and accepted October 7, 2013. doi: 10.1111/trf.12603 © 2014 AABB TRANSFUSION 2014;54:1634-1641.
BACTX ASSAY FOR BACTERIA IN PLTs
accomplished by culture using broth cultures of a sample of apheresis PLTs and, where pooled at blood centers, pooled whole blood–derived PLTs taken approximately 24 hours postdonation, associated with sample reading 36 to 48 hours postdonation and release of units to transfusion centers if culture results are negative at this time.9 Testing of whole blood–derived PLT concentrates pooled immediately before release, which was previously allowed using simple but insensitive and nonspecific pH or glucose measurements, is now required to be performed using specific Food and Drug Administration (FDA)-approved bacterial assays (AABB Standard 5.1.5.1.1, effective January 2011).10 These measures have helped lower the number of fatalities due to bacterial contamination of PLT products reported to the FDA from approximately seven per year to two to three per year, predominantly associated with reductions in deaths due to Gram-negative bacterial species.11 Detection of bacterial contamination of PLT products by bacterial culture as currently performed is limited by the low bacterial load present at the time of collection.12 Initial studies of bacterial contamination were modeled on bacterial loads of 0.154 colony-forming units (CFUs)/ mL, or approximately 50 CFUs per apheresis product, where detection by culture of 4- to 8-mL aliquots was likely. However, later studies suggested that loads 1/10 of this were not uncommon in many units.2,7,13,14 Since industry standards include the sampling of approximately 8 mL of PLT concentrates or approximately 2% of the product volume, it has been estimated by Poisson sampling analysis that current culture techniques would detect less than 50% of contaminated units.15,16 Other limitations of early culture include the variable rates of bacterial growth at low bacterial loads and the fact that only anaerobic culture (not now a blood center standard) detected low levels of common skin contaminant Gram-positive cocci.15,16 All these limitations of the current culture approach toward minimizing the hazard of bacterial contamination of PLT concentrates are reflected in a recent study suggesting that less than 50% of bacterially contaminated apheresis units are detected by early culture.2 For all these reasons a sensitive method for testing apheresis and whole blood–derived PLT units at time of issue is needed. Two methods are now FDA-approved and meet regulatory requirements—the PGD test (Verax Biomedical, Marlborough, MA), and the BacTx assay (Immunetics, Inc., Boston, MA), which are both cleared for quality control testing of apheresis units and pools of up to 6 whole blood–derived units.17,18 The Immunetics BacTx assay, intended for use as a qualitative test for bacterial contamination of PLT products, detects peptidoglycan, a universal component of cell walls of both Grampositive and Gram-negative bacteria.19 The assay is based on an insect innate immune system bacterial defense mechanism, in which peptidoglycan-binding proteins
from insect hemolymph trigger a prophenoloxidase serine protease cascade. The phenoloxidase enzyme activity converts catechols to quinones, which are complexed with MBTH, a soluble chromogen that turns red upon quinone binding, generating measurable absorbance at 500 nm. Absorbance is monitored by a dedicated photometer over a fixed time period. This study examined the analytical sensitivity and time to detection of the BacTx assay using PLT units inoculated with 10 representative bacterial species. Specificity and reproducibility across test kit lots were also studied. All testing was done at two study sites, Cleveland, Ohio, and Manhasset, New York.
MATERIALS AND METHODS PLT units All testing was performed at two study sites using indate, leukoreduced, apheresis PLT (LRAP) units and leukoreduced, whole blood–derived PLT (LRWBDP) units acquired from blood centers. LRWBDP minipools were prepared as needed by pooling 2-mL volumes from LRWBDP units, using sterile transfer methods, with concurrent culture screening to verify sterility of units.
Bacterial strains Bacteria used in the analytical sensitivity, time to detection, and reproducibility assays included the following 10 bacterial species (ATCC number): aerobic or facultative Gram-positive species, Staphylococcus aureus (27217), Staphylococcus epidermidis (49134), Bacillus cereus (11778), and Streptococcus agalactiae (12386); aerobic or facultative Gram-negative species, Serratia marcescens (43862), Pseudomonas aeruginosa (27853), Escherichia coli (25922), and Klebsiella oxytoca (43863); and anaerobic Gram-positive species, Clostridium perfringens (3629) and Propionibacterium acnes (11827).
BacTx assay The BacTx assay (Immunetics, Inc.) for bacterial peptidoglycan is provided as a kit comprising all necessary reagents to carry out multiple tests. A photometer is provided as a required accessory and has positions for up to eight tests to be run simultaneously. The photometer is attached to a dedicated personal computer that records information on specimens being tested and results obtained from the photometer. Other equipment required includes a microcentrifuge capable of 14,000 to 20,000 × g, a variable-speed mixer, and micropipettors capable of delivering up to 1000-μL volumes. To perform the BacTx assay with LRWBDP, sample volumes of 0.5 mL are sterilely obtained from PLT pools to be tested and placed in Volume 54, June 2014
TRANSFUSION
1635
HEATON ET AL.
2-mL microcentrifuge tubes containing 1.0 mL of lysis buffer. Microcentrifuge tubes are then immediately capped, inverted three times to mix, and centrifuged at 18,000 × g for 3 minutes. For LRAP, 1.0 mL of PLT are combined with 0.75 mL of lysis buffer and mixed vigorously before centrifugation. For all PLT types, the supernatant is then decanted into a waste container and any remaining droplets are removed by tapping tubes against the rim of the waste container. The remaining pellets are then resuspended in 0.5 mL of extraction reagent by pipetting up and down 10 times. Volumes of 0.5 mL of neutralization reagent are placed in 1.5-mL microcentrifuge tubes and the resuspended pellets are transferred to these tubes, which are then capped and inverted three times to mix. Volumes of 0.3 mL are then transferred to BacTx reaction tubes, which are then briefly mixed vigorously and placed in the BacTx reader. The BacTx reader cycle is started from the attached computer and reactions in the BacTx reaction tubes are monitored by the BacTx reader software for changes in color (absorbance) over 30 minutes. An increase in absorbance within this time period above a predetermined cutoff is interpreted by the software as a positive or “fail” result, indicating bacterial contamination. Otherwise, a negative or “pass” result is delivered at the end of the time period. Due to nonlinear signal amplification the signal is not directly related to input bacterial concentration, but the time to exceed the assay cutoff is affected by bacterial input concentration.19 Results are stored electronically in an assay log file, which can also be printed and exported to a laboratory information system. The turnaround time for testing a batch of up to eight PLT samples is less than 45 minutes.
Analytical sensitivity The analytical sensitivity of the BacTx assay was determined in LRAP and freshly prepared 6-unit LRWBDP minipools in tubes using bacterial suspensions prepared at targeted counts. Bacterial species were grown overnight on aerobic or anaerobic blood agar plates and then inoculated into appropriate broth (tryptic soy broth or brain heart infusion broth for aerobes; anaerobic Brucella broth for anaerobes) and incubated at 35°C for several hours until target optical density (OD) values were obtained. Bacterial suspensions were then diluted in phosphatebuffered saline (PBS) to defined bacterial load levels calibrated by OD against a standard curve to allow initial dilution in LRAP and LRWBDP pools (pools prepared from 2-mL aliquots from 1 inoculated and 5 uninoculated units to form 6-unit minipools of 12 mL) at defined bacterial titers. These targeted bacterial titers in LRAP and LRWBDP pools were 1 × 103, 5 × 103, 1 × 104, 2 × 104, 4 × 104, and 8 × 104 CFUs/mL. Ten replicates of LRAP and LRWBDP minipools containing each species and concentration 1636
TRANSFUSION Volume 54, June 2014
were tested by BacTx assay and interpreted as detected at the lowest bacterial titer at which all 10 replicates tested positive. Quantitative bacterial culture was performed on samples from each PLT or pool by plating 0.1 mL of serial 10-fold dilutions in duplicate onto blood agar plates to determine actual bacterial titers in the PLT or pool at the time of testing.
Time to detection This was evaluated by inoculating very low titers (target, 1-5 CFUs/mL) of each of the 10 bacterial species into LRAP and individual LRWBDP units. Bacterial species were grown overnight on aerobic or anaerobic blood agar plates and inoculated into appropriate broth and incubated at 35°C to specific levels defined by OD as described above. After dilution in PBS to defined bacterial load levels, 4 LRAP and 4 LRWBDP units were inoculated (two with 0.1 mL and two with 0.3 mL) and sampled by plate culture incubated overnight to determine actual bacterial inocula. An additional unit was inoculated with saline to serve as a negative control. Inoculated units were incubated at 22°C in a PLT incubator-rocker for 7 days. On Day 1, (t = 24 hr postinoculation), the bacterial inocula were determined from growth plates, and two of each set of 4 inoculated units were chosen for further study based on the bacterial inocula being within the targeted range of 1 to 5 CFUs/mL. Aliquots were removed from the inoculated units selected for study and 0.1 mL plated onto blood agar plates. If growth was detected on these plates after incubation for 24 hours (t = 48 hr), indicating that bacterial growth had occurred, one of the inoculated units was selected for further study and was incubated for a total of 7 days (168 hr). On Day 2 (t = 48 hr) and Day 7 (t = 168 hr), an aliquot from inoculated units in the case of LRAP units, or from a LRWBDP pool prepared from an inoculated LRWBDP unit pooled with aliquots from 5 fresh sterile LRWBDP units, was tested by BacTx assay in replicates of 10. In addition the saline-inoculated unit was tested by BacTx assay in triplicate at each time point. If fewer than 10 replicates from a unit were positive at t = 48 hours the unit was retested at t = 72 hours. Quantitative bacterial counts were done by plate culture at t = 0 on inocula and on the inoculated units or pools on each day of BacTx assay testing. The bacterial species used were identified independently from the initial inoculum and after recovery on Day 7 to assure that the species detected were those inoculated.
Specificity Specificity testing was performed on uninoculated LRAP and uninoculated 6-unit LRWBDP minipools. To confirm sterility LRWBDP pools and LRAP units were cultured by plate culture in triplicate on aerobic blood agar plates,
BACTX ASSAY FOR BACTERIA IN PLTs
TABLE 1. Analytical sensitivity of the BacTx assay in LRAP units and LRWBDP minipools containing bacteria added at titers ranging from 1 × 103 to 8 × 104 CFUs/mL, with actual bacterial titers determined by quantitative plate culture at time of testing*
Bacterial species B. cereus C. perfringens E. coli K. oxytoca P. acnes P. aeruginosa S. marcescens S. aureus S. epidermidis S. agalactiae
Lowest bacterial titer (CFUs/mL) at which all 10 samples were positive by BacTx assay by PLT type and test site LRAP LRWBDP Site 1 Site 2 Site 1 Site 2 1.9 × 103 1.3 × 103 1.7 × 103 1.4 × 103 9.4 × 102 4.8 × 103 2.8 × 103 4.5 × 103 7.6 × 104 3.3 × 104 5.1 × 103 8.7 × 103 1.6 × 104 7.3 × 103 6.8 × 103 9.9 × 103 5.0 × 103 8.5 × 103 7.2 × 103 1.1 × 103 2.7 × 104 2.0 × 104 9.6 × 103 5.0 × 104 4.2 × 103 5.3 × 103 5.8 × 104 6.7 × 103 2.2 × 103 1.1 × 103 2.1 × 103 4.0 × 103 1.3 × 103 6.3 × 102 2.0 × 103 2.4 × 103 4.5 × 103 3.3 × 103 3.6 × 103 2.7 × 104
* LRWBDP minipools were prepared from one LRWBDP unit containing bacteria pooled with 5 uninoculated LRWBDP units. Ten replicates of the BacTx assay were performed for each bacterial species and concentration and analytic sensitivity defined as the lowest bacterial titer at which all 10 samples were positive by BacTx assay.
using 0.1-mL volumes. LRAP units were also cultured by an automated bacterial detection system (BacT/ALERT, bioMérieux, Durham, NC) in BPA and BPN bottles, using 4-mL aliquots per bottle. For LRWBDP, on each day of testing at each site, 8 individual PLT units were used to prepare 24 unique minipools containing 1-mL aliquots from 6 different units, organized so that each minipool had a set of six unique constituents. A different set of 8 PLT units was used for each of the nine testing iterations at each site, with a total of 144 LRWBDP units used and 432 unique minipools (216 at each site) prepared. On each day of testing the 24 unique minipools were tested in three groups of eight using three different BacTx kit lots at the two testing centers (three runs of 24 pools per lot per site using one lot per day). LRAP units were tested with five BacTx kit lots. For LRAP, a retest algorithm was used to determine assay reactivity. LRAP units that were initially reactive in the BacTx assay were retested in duplicate. If either retest was positive, the test result was confirmed as reactive. If neither retest was reactive, the unit was designated as nonreactive.
Reproducibility Interlot and intersite reproducibility were tested using frozen aliquots of 10 bacterial suspensions in saline of the 10 bacterial species used above and a saline negative control. The aliquots contained bacteria at titers that were 0.5 to 1.5 log CFUs/mL higher than the sensitivity determined for a given strain in the analytical sensitivity portion of the study and were frozen in tubes in 100-μL volumes. For LRAP, a test panel was used for testing at the two sites using three lots of BacTx assay kits on 3 different days. The test panel consisted of 10 bacterial members and one negative member. The sterility of LRAP used in
the study was confirmed by either BacT/ALERT culture or agar plate culture. Similarly, for LRWBDP, three different BacTx kit lots were used at both sites on 3 different days. On each testing day, four unique 18-mL PLT pools were prepared using a combination of 3-mL aliquots from 6 of 7 individual LRWBDP units. Aliquots of 0.1 mL from each pool were placed on blood agar plates and incubated aerobically to assure sterility of the PLT pools. Aliquots of 0.5 mL from each LRWBDP pool or 1.0 mL from each LRAP unit were added to each of the 10 thawed bacterial suspension aliquots and a saline control, and then tested by BacTx assay.
RESULTS Analytical sensitivity The analytic sensitivity of the BacTx assay on LRAP and 6-unit LRWBDP minipools inoculated with the 10 representative PLT bacterial contaminants studied ranged from 6.3 × 102 CFUs/mL for S. epidermidis to 7.6 × 104 CFUs/ mL for E. coli (Table 1). Overall, five of the bacterial species were detected at less than 104 CFUs/mL in both PLT types at both sites, while the remaining species were detected at not more than 7.6 × 104 CFUs/mL.
Time to detection LRAP units were inoculated with 1.3 to 5.3 CFUs/mL and LRWBDP units with 0.7 to 5.0 CFUs/mL. Time to detection in LRAP and 6-unit LRWBDP minipools, prepared at each time of testing from 1 inoculated and 5 sterile units, is shown in Table 2. Bacterial detection in all 10 replicates was found at 48 hours for 28 of the 32 challenges with the eight aerobic or facultative challenge bacterial species, with detection in all 10 replicates in the remaining four Volume 54, June 2014
TRANSFUSION
1637
HEATON ET AL.
TABLE 2. Results of time to detection studies after inoculation of LRAP and LRWBDP units with aerobic bacterial species* Number of replicates (out of 10) detected by BacTx assay by PLT type and time point LRWBDP Bacterial species E. coli P. aeruginosa K. oxytoca S. marcescens B. cereus S. aureus S. epidermidis S. agalactiae
Site number 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2
48 hr 10 10 10 10 10 10 10 10 10 10 10 10 9 10 10 10
LRAP 72 hr —† — — — — — — — — — — — 10 — — —
48 hr 10 10 0 0 10 10 10 10 10 10 10 10 10 10 10 0
72 hr — — 10 10 — — — — — — — — — — — 10
* LRAP units were inoculated with 1.3 to 5.3 CFUs/mL bacteria at Time 0. LRWBDP units were inoculated with 0.7 to 5.0 CFUs/mL of bacteria at Time 0, with inoculated units pooled with 5 uninoculated LRWBDP units at time of testing. Ten replicates were tested by BacTx assay at 48 hours. If one or more of the 10 aliquots was negative at 48 hours, testing was repeated at 72 hours. † Not tested at 72 hours as all 10 replicates were positive at 48 hours.
challenges at 72 hours. The challenge organisms in these four cases where detection occurred at 72 hours were P. aeruginosa in LRAP units at both sites; S. epidermidis in a LRWBDP minipool at one site, where bacterial contamination was detected by the BacTx assay in nine of 10 replicates at 48 hours at a bacterial titer of 6.6 × 102 CFUs/mL; and S. agalactiae in LRAP at one site (Table 2). Bacterial titers in pools at time of detection of all 10 replicates ranged from 8.7 × 102 CFUs/mL for S. agalactiae to 6.4 × 108 CFUs/mL for E. coli. The two challenge anaerobic species studied, P. acnes and C. perfringens, failed to grow in inoculated PLT units and remained undetectable at all time points tested by both plate culture and BacTx assay.
TABLE 3. Specificity of the BacTx assay in sterile LRAP and LRWBDP at the two test sites BacTx assay result Negative Initially reactive Repeat reactive
LRAP units (n = 505) 501 4 1
LRWBDP pools (n = 432)* 431 1 ND
* Prepared from 144 LRWBDP units. ND = not determined.
tests performed had to be retested due to an aborted assay run.
Reproducibility Specificity A total of 432 unique LRWBDP pools, derived from 144 LRWBDP units, were tested (216 at each site) in three groups of 8 on 9 separate days using three different BacTx kit lots. There was no bacterial growth associated with any of these pools. Of these 432 pools, 431 tested correctly as negative, for a test specificity of 99.8% (the one false positive was not retested) (Table 3). For LRAP, a total of 505 units were tested (96 units at Site 1 and 409 at Site 2) with five BacTx kit lots. Of the 505 units tested, 501 were negative on initial testing, and four were initially reactive. Upon retesting, only one of the seven initially reactive samples was reactive (Table 3). Therefore, 504 of 505 units were initially negative or nonreactive on repeat testing for a test specificity of 99.8% for LRAP. One of the 937 specificity 1638
TRANSFUSION Volume 54, June 2014
A total of 528 reproducibility tests were run on LRAP and LRWBDP pools (132 assays with each PLT type at each site), using three different BacTx kit lots per PLT type (six kit lots in total) on 3 different days on the 11 PLT specimens prepared. In all 480 cases where bacterial suspensions had been added, the BacTx assays on the PLT specimens yielded positive results, while the 24 LRAP and 24 LRWBDP pool specimens inoculated with saline yielded negative results (Table 4). In total, all 528 BacTx assays yielded correct results, for 100% reproducibility.
DISCUSSION There is considerable interest in addressing the continuing problem of bacterial contamination of PLT products.20
BACTX ASSAY FOR BACTERIA IN PLTs
TABLE 4. Reproducibility of the BacTx assay using three BacTx assay kit lots at the two test sites Lot 1 2 3
LRAP 88/88* 88/88 88/88
LRWBDP 88/88 88/88 88/88
* Results are expressed as number of tests with expected result/number of tests performed.
A 1-day public conference on this topic was sponsored by AABB in July 2012 with the objectives of discussing the residual risk of bacterial contamination in apheresis PLTs, describing the use of secondary bacterial testing of PLT components using point-of-issue tests, and hearing practical experiences from transfusion services using these tests with whole blood–derived PLTs.21 Presenters at this meeting noted that, although there is no FDA requirement for testing of PLT products for bacterial contamination, several products have been approved by the FDA to decrease the risk of bacterial contamination, including diversion pouches and culture systems such as BacT/ ALERT and eBDS, as well as point-of-issue tests such as the PLT PGD test and the BacTx assay system. On September 21, 2012, the Blood Products Advisory Committee of the FDA met to discuss considerations for strategies to further reduce the risk of bacterial contamination in PLTs.22,23 The committee voted unanimously in favor of additional measures being necessary to decrease the current risk of transfusion of bacterially contaminated PLT products. The committee voted unanimously against reducing PLT product shelf life from 5 to 4 days as being insufficient to decrease the risk of transfusion-associated septic reactions and to obviate the need for additional testing. The committee also voted in favor of a strategy of culturing PLTs after the first 24 hours of storage, as currently practiced, and then retesting on Day 4 or Day 5 just once with a rapid test on the day of transfusion; however, the vote was split on applying this strategy to retesting on Day 3. The FDA is now considering issuing guidance on additional testing requirements for PLTs near time of issue. The BacTx assay is a simple rapid test for bacterial contamination in a tube format that is suitable for near time of issue use and can be performed in under 1 hour. In our studies the assay successfully detected 10 representative bacterial species inoculated into representative PLT concentrates or pools, with detection at concentrations ranging from 6.2 × 102 to 7.6 × 104 CFUs/mL in LRAP and pooled LRWBDP. The threshold of bacterial detection of the assay was therefore below the bacterial load of more than 105 CFUs/mL reported to be associated with septic transfusion reactions.24 Compared with the analytic sensitivity values for a similar range of species of approximately
104 to 106 CFUs/mL reported for the PGD test, the BacTx assay was more sensitive overall, by more than an order of magnitude in several instances.18,25 For example, in LRAP, the BacTx assay was more than 100-fold more sensitive than the PGD test for detection of S. marcescens (5.3 × 103 CFUs/mL vs. 8.2 × 105 CFUs/mL) and more than 10-fold more sensitive for detection of S. agalactiae (4.5 × 103 CFUs/mL vs. 5.5 × 104 CFUs/mL) and C. perfringens (4.8 × 103 CFUs/mL vs. 8.9 × 104 CFUs/ mL).18,25 Of the 10 bacterial strains used in the current work, only one, S. marcescens ATCC 43862, was the same strain of this species used in evaluations of analytic sensitivity with the PGD test. Comparisons between the analytic sensitivity values found in this and the PGD test studies are therefore confounded by the use of predominantly different strains that may have distinct growth characteristics, making direct comparisons less precise. The bacterial species studied included four each of aerobic or facultative Gram-positive and Gram-negative species, representing the more significant organisms reported to be associated with posttransfusion bacterial reactions.11,20,26 In addition two anaerobic Gram-positive species also associated with transfusion reactions were detected, although the anaerobic bacteria failed to grow in the PLTs, so time to detection could not be evaluated. In the PLT units spiked with inocula between 0.7 and 5.3 CFUs/mL, 28 of the 32 challenges with the eight aerobic species were detected at 48 hours, with the remaining four challenges detected at 72 hours. These findings are similar to those found in time to detection studies using the PGD test,18 in which three of the strains studied, S. marcescens ATCC 43862, S. aureus ATCC 27217, and S. epidermidis ATCC 49134 were the same as those used in our time to detection evaluations of the BacTx assay. Publications based on the implications of time to culture positivity have suggested that, while the median number of bacteria initially contaminating PLT units may be as high as 50 CFUs/product, very low loads of approximately 5 CFUs/product are more common.2,14,15,27 Additionally, bacteria may not start growing immediately and can have an extended lag phase. Under such circumstances the time testing is performed, the length of time needed to generate a result and the limit of detection of the test method are important variables. Although recent reports based on bacterial growth curves with culture using time to positive modeling have suggested that inoculation of larger samples into culture bottles might increase the probability of detection of contamination, these studies do not address the problems posed by delayed bacterial growth in PLT with low inocula. The relatively low assay sensitivity threshold and short time to detection of contamination reported for the BacTx assay in the current study suggest that application of this assay as a point-of-release test might confer an additional level of transfusion safety.2,14,15,28 Volume 54, June 2014
TRANSFUSION
1639
HEATON ET AL.
Specificity with multiple lots and technical operators was high, at more than 99.8%, with a very low invalid run rate (one of 937 tests). Reproducibility across different lots and with different operators was likewise excellent at 100%, and equipment failures were minimal. As such the BacTx assay appears well suited for use in a routine laboratory environment. In this evaluation, the BacTx assay demonstrated the highest levels of sensitivity and specificity reported for a cleared commercial rapid test for detection of bacterial contamination in PLTs. The assay would thus serve the need for point-of-release bacterial screening of apheresis and pooled random-donor PLT concentrates at 3 or more days of room temperature storage to prevent transfusion of PLT units and pools contaminated with clinically significant levels of bacteria.3,27 The use of this test to extend the shelf life of PLTs beyond the current 5-day period should also be explored. Clinical trials are warranted to determine the value of this product in detecting bacterial contamination of PLT units administered to patients. ACKNOWLEDGMENTS The authors are grateful for technical support from Munira Mulani, MT. The authors acknowledge the assistance of the Immunetics Support Team, Andrew Han, PhD, Neil Krueger, PhD, and Richard Pinkowitz.
CONFLICT OF INTEREST All authors except MRJ, RAY, and WAH have declared no other conflict of interest or financial involvement with this article. MRJ has received research support and/or honoraria from Verax, Pall, Gambro, Hemosystem, Immunetics, Genprime, and Charles River Labs and has been a consultant for BioSense Technologies and Lynntech, Inc. RAY has received honoraria and/or consulting fees from Immunetics, Verax, Pall, Gambro, Hemosystem, Fenwal, and Genprime. WAH has received research support/honoraria support from Verax, Immunetics, Fenwal, Beckman Coulter, Novartis Diagnostics, TerumoBCT, and Haemeonetics.
REFERENCES 1. Dodd R, Kurt Roth W, Ashford P, et al. Transfusion medicine and safety. Biologicals 2009;37:62-70. 2. Eder AF, Kennedy JM, Dy BA, et al. Limiting and detecting bacterial contamination of apheresis platelets: inlet-line diversion and increased culture volume improve component safety. Transfusion 2009;49:1554-63. 3. Jacobs MR, Smith D, Heaton WA, et al. Detection of bacterial contamination in prestorage culture-negative apheresis platelets on day of issue with the Pan Genera Detection test. Transfusion 2011;51:2573-82. 4. Zou S, Dorsey KA, Notari EP, et al. Prevalence, incidence, and residual risk of human immunodeficiency virus and 1640
TRANSFUSION Volume 54, June 2014
hepatitis C virus infections among United States blood donors since the introduction of nucleic acid testing. Transfusion 2010;50:1495-504. 5. Zou S, Musavi F, Notari EP, et al. Prevalence, incidence, and residual risk of major blood-borne infections among apheresis collections to the American Red Cross Blood Services, 2004 through 2008. Transfusion 2010;50: 1487-94. 6. Zou S, Stramer SL, Notari EP, et al. Current incidence and residual risk of hepatitis B infection among blood donors in the United States. Transfusion 2009;49:1609-20. 7. Benjamin RJ, Wagner SJ. The residual risk of sepsis: modeling the effect of concentration on bacterial detection in two-bottle culture systems and an estimation of falsenegative culture rates. Transfusion 2007;47:1381-9. 8. American Association of Blood Banks. Standards for blood banks and transfusion services. 22nd ed. Standard 5.1.5.1. Association Bulletin 03-12, appendix I. 2003. 9. Brecher ME, Jacobs MR, Katz LM, et al. Survey of methods used to detect bacterial contamination of platelet products in the United States in 2011. Transfusion 2013;53:911-8. 10. American Association of Blood Banks. Standards for blood banks and transfusion services. 27th ed. Summary of significant changes. 2011 [cited 2013 Oct 4]. Available from: http://www.aabb.org/sa/standards/Documents/ sigchngstds27.pdf 11. FDA. Fatalities reported to FDA following blood collection and transfusion: annual summary for fiscal year 2012. 2013. [cited 2013 Oct 4]. Available from: http://www.fda .gov/BiologicsBloodVaccines/SafetyAvailability/ ReportaProblem/TransfusionDonationFatalities/ ucm346639.htm 12. Eder AF, Kennedy JM, Dy BA, et al. Bacterial screening of apheresis platelets and the residual risk of septic transfusion reactions: the American Red Cross experience (20042006). Transfusion 2007;47:1134-42. 13. Murphy WG, Foley M, Doherty C, et al. Screening platelet concentrates for bacterial contamination: low numbers of bacteria and slow growth in contaminated units mandate an alternative approach to product safety. Vox Sang 2008; 95:13-9. 14. Pearce S, Rowe GP, Field SP. Screening of platelets for bacterial contamination at the Welsh Blood Service. Transfus Med 2011;21:25-32. 15. Benjamin RJ, Dy B, Perez J, et al. Bacterial culture of apheresis platelets: a mathematical model of the residual rate of contamination based on unconfirmed positive results. Vox Sang 2014;106:23-30. 16. Tomasulo PA, Wagner SJ. Predicting improvement in detection of bacteria in apheresis platelets by maintaining constant component sampling proportion. Transfusion 2013; 53:835-42. 17. Immunetics. BacTx bacterial detection system. 2013. [cited 2013 Oct 4]. Available from: http://www.immunetics.com/ bactx.htm
BACTX ASSAY FOR BACTERIA IN PLTs
18. Verax Biomedical. Platelet PGD Test. 2013. [cited 2013 Oct 4]. Available from: http://www.veraxbiomedical.com/ products/platelet-pgd-test.asp 19. Kovalenko V, Levin A, Beausang L. Rapid peptidogycan-
23. Epstein JS, Vostal JG. FDA contributions to reduction of bacterial contamination in platelet products within the United States. Transfusion 2013;53:232-3. 24. Jacobs MR, Good CE, Lazarus HM, et al. Relationship
based assay for detection of bacterial contamination of
between bacterial load, species virulence, and transfusion
platelets. US Patent 7,598,054. 2009. [cited 2013 Oct 4]. Available from: http://www.google.com/patents/
reaction with transfusion of bacterially contaminated platelets. Clin Infect Dis 2008;46:1214-20.
US7598054 20. Brecher ME, Blajchman MA, Yomtovian R, et al. Addressing the risk of bacterial contamination of platelets within the United States: a history to help illuminate the future. Transfusion 2013;53:221-31. 21. American Association of Blood Banks. Public Conference— Secondary Bacterial Screening of Platelet Components, 7/17/12, Bethesda, MD. 2012. [cited 2013 Oct 4]. Available from: http://www.aabb.org/events/misc/Pages/public -conference.aspx 22. FDA. Meeting of the Blood Products Advisory Committee to discuss considerations for strategies to further reduce the risk of bacterial contamination in platelets, Rockville, MD. 2012. [cited 2013 Oct 4]. Available from: http://
25. Vollmer T, Hinse D, Schottstedt V, et al. Inter-laboratory comparison of different rapid methods for the detection of bacterial contamination in platelet concentrates. Vox Sang 2012;103:1-9. 26. Brecher ME, Means N, Jere CS, et al. Evaluation of an automated culture system for detecting bacterial contamination of platelets: an analysis with 15 contaminating organisms. Transfusion 2001;41:477-82. 27. Kuehnert MJ, Roth VR, Haley NR, et al. Transfusiontransmitted bacterial infection in the United States, 1998 through 2000. Transfusion 2001;41:1493-9. 28. Murphy WG. Pathogen reduction: state of reflection in Ireland. Transfus Clin Biol 2011;18:488-90.
www.fda.gov/AdvisoryCommittees/Calendar/ ucm313863.htm
Volume 54, June 2014
TRANSFUSION
1641
