Validation of new assay for bacterial detection of RDPs

March 1, 2011

A Food and Drug Administration (FDA) mandate and AABB standards1 rendered pH and glucose methodologies for detecting bacteria in random donor platelets (RDPs) obsolete as of Jan. 31, 2011. The pan genera detection, or PGD, assay received 501(k) clearance from the FDA for bacterial detection in both apheresis and RDPs in 2007. The PGD technology detects the presence of conserved antigens lipotechoic acid (LTA) and lipopolysaccharide (LPS) on Gram-positive (GP) and Gram-negative (GN) bacteria, respectively. It is a CLIA non-waived qualitative test of moderate complexity and can detect bacteria as low as 103 to 105 CFU/mL. Up to six RDPs may be tested in a pooled sample within four hours of transfusion. If the pooled sample is non-reactive, the platelets in the pool may be transfused, satisfying AABB Standard 5.1.5.1.1.

Time to perform the test constitutes approximately 20 minutes of technologist time with an additional minimum of 20 minutes (maximum of 60 minutes) for interpretation. Briefly, the pooled sample is mixed with reagents, centrifuged, and manipulated to fragment the platelet pellet. The sample is then vortexed and transferred to the test device where lateral-flow technology facilitates binding of bacteria, if present, to LTA and LPS targets on the test strip. Reactivity is clearly visualized by pink bands in GP and/or GN windows. Internal controls are also incorporated into the test device as a further check of test validity. External controls need only be run once per kit.

The example transfusion service used in this study procures blood products for a 296-bed not-for-profit hospital. The transfusion service utilizes few RDPs to satisfy patient needs; single-donor platelets are usually available from the blood supplier. The decision to validate and implement the PGD assay was made to satisfy patient needs in times of platelet shortages where RDPs may be the only product available.

Validation process: The validation process started with communication with and approval by the medical director and supervisor of the transfusion service and laboratory director. A representative from the vendor was contacted to acquire all pertinent information on the assay, i.e. pricing, equipment needs, preliminary studies. Information in the literature indicated the PGD assay was tested with a variety of GP and GN organisms with acceptable sensitivity and specificity results. Comparative studies were also performed between PGD assay developed by Verax Corp. (Worcester, MA) and BacT/ALERT system demonstrating 10 isolates detected by the BacT/ALERT system were reactive in the PGD assay.2 The representative also recommended a mixed titer bacteria panel for validation studies (Zeptometrix Corp, Buffalo, NY).

Methods: The study design involved testing RDPs from the blood supplier using current methodology of pH and glucose, PGD assay, and culture. The number of samples tested was 20. The RDPs used for validation were tested within 12 hours of receipt from the blood supplier and were in date at the time of testing. The ABO/Rh type of the RDPs was not a consideration for testing. All types were acceptable.

There were three resources of samples used in this study:

  1. RDPs from the blood supplier not containing any bacterial contamination (n=6),
  2. samples prepared using the mixed titer bacteria panel (n=12), and
  3. samples from the College of American Pathologists (CAP) surveys (n=2).

The platelet sample from blood-supplier RDPs was procured by stripping tubing from the RDP tail and pipetting 500 uL into a microcentrifuge tube provided with kit. Eight drops of Reagent 1 were then added to the sample, and the mixture was centrifuged between 9,000 and 11,000 RCF for five minutes. After centrifugation, the supernatant was decanted and eight drops of Reagent 2 were added to platelet pellet in a microcentrifuge tube. The pellet was then dislodged into small fragments using a disposable pipet provided with kit. Four drops of Reagent 3 were added, and the mixture was vortexed. The mixture was then added to the center well of the test device. A humidifying cover was placed over the test device. Interpretation was made between 20 and 60 minutes. The mixed titer bacteria panel samples were prepared by adding 1 mL of RDP to a cryovial containing bacteria. After mixing, the PGD assay was performed as described above. Samples from CAP surveys were prepared by adding 1 mL of diluent to lyophilized pellet and transferring this mixture to simulated platelet vial. Glucose and pH testing was done by adding a drop of the platelet sample to urinalysis test strips and interpreting color change in one minute. A pH of

Results: Twelve of the samples were known to contain bacteria, and all 12 yielded reactive results on the PGD assay. The identification of the bacteria corresponded to reactivity in respective GP and GN windows of the PGD test device, (i.e., samples containing Escherichiacoli revealed a pink band in the GN window. The two known CAP positive samples yielded positive cultures (Bacillus sp, Staphylococcus sp.) and reactive PGD. The 10 known positives using the mixed titer bacteria panel yielded reactive PGD in all 10 samples, but only four were positive on culture. A possible explanation of why the CAP-positive samples yielded reactive PGD and confirmatory culture, and the mixed titer bacteria samples confirmed (in culture) just 40% of known positives is the latter contained 0.1% of sodium azide, which may inhibit bacteria growth but still be reactive in PGD assay. Organisms present in the mixed titer bacteria panel included Bacillus thuringienis, E coli (low and high concentrations), Staphylococcus aureus (low and high concentrations), Enterobacter aerogenes, Pseudomonas aeruginosa, Streptococcus dysgalactiae, Klebsiella pneumoniae, S epidermidis. Organisms which appeared to be resistant to the sodium azide preservative were Bacillus thuringienis, S aureus (low and high), and S dysgalactiae. The RDPs from the blood supplier that were not modified or spiked with bacteria yielded non-reactive PGD and negative culture. The pH and glucose results were variable. Thirteen of 20 samples (65%) were abnormal. Sensitivity for culture and PGD assay was 100%, specificity was 57%, and positive-predictive value (PPV) was 50%. Sensitivity for culture and pH/glucose was 83%, specificity was 42%, and PPV was 36%.

Discussion

Information learned throughout the validation process for the PGD assay included ease of running the assay, time commitment by technologists, clarity of discerning a reactive from a non-reactive test, and comparisons with current and confirmatory tests. The PGD assay is very easy to perform. Approximately 20 technologists/technicians have been trained since validation without any problems. Most of the technologist time will be spend on procuring the pooled sample from up to six RDPs and isolating the platelet pellet. Once the sample is loaded onto the hand-held test device, the technologist can essentially walk away and come back 20 minutes later. As many as two bands can be seen in a reactive test and are very clear. Non-reactive tests simply contain no bands in the GP and GN windows. Regarding the internal controls wells, an inkling of blue color on both ends of the test device signifies interpretation can be made. The internal control pads do not need to be covered completely in blue. The PGD kits have a shelf life of about four months. Kits containing 20 test devices and 100 test devices are available. The one capital equipment item blood banks may not have access to is a microcentrifuge with 9,000 to 11,000 RCF capabilities.

All samples which did not contain bacteria yielded non-reactive results on PGD. In contrast, all samples containing bacteria yielded reactive PGD results. The discrepancy in the culture results of samples containing bacteria are explained by the presence of sodium azide in the mixed titer bacteria panel samples. Blood banks who have not yet performed validation might consider using positive samples containing no sodium azide. Sensitivity, specificity, and PPV results were higher between the new method and culture than with the current method and culture. A study conducted at the University of Pittsburg3 also utilized the mixed titer bacteria panel for validation and found that those samples containing bacteria yielded a reactive PGD test, and samples containing no bacteria yielded a non-reactive PGD test. Cultures were not performed. Vollmer, et al4 spiked apheresis platelets with various organisms and analyzed time to detection in PGD assay and minimum concentration to detection. The results for Gram-positive strains matched specifications from the manufacturer; however select strains of E coli and K pneumoniae demonstrated higher concentrations for detection than that documented from the manufacturer as well as longer times for initial detection. This finding is suggestive of possible false-negative samples and future studies should focus on these two organisms.

The example transfusion service will perform PGD testing on a pooled sample of RDPs. If the pooled sample is non-reactive, the platelets in the pool will then be physically pooled. If the pooled sample is reactive, the platelets in the pool will be tested individually by the PGD assay. This process is being followed to ensure only non-reactive RDPs are pooled and to avoid wasting platelet products. Given the variability of pH and glucose testing where bacterial detection relies on various shades of color, which may not always match the color legend and a decreased pH is not specific for bacterial contamination in the presence of metabolically active white blood cells and platelets,3 the PGD assay offers direct bacterial membrane capture by LTS and LPA antibodies in a GP and GN split window testing device.

Virginia C. Hughes, PhD(ABD), SBB(ASCP), is section chief of Transfusion Services at the Shore Memorial Hospital in Somers Point, New Jersey

 

Acknowledgement. The author acknowledges the Lee Collins-McStravick for microbiology technical assistance and equipment usage.

_____________

References

  1. Price T. Standards for Blood Banks and Transfusion Services ed. 26, AABB, 2009.
  2. Platelet PGD Test product insert.
  3. YazerMH, Stapor DS, Triulzi DJ. Use of the RQI test for bacterial screening of whole blood platelets. Am J Clin Pathol. 2010;133:564-568.
  4. Vollmer T, Hinse D, Kleesick K, Dreier J. The Pan Genera Detection Immunoassay: a novel point of issue method for detection of bacterial contamination in platelet concentrates. J Clin Micro. 2010;48:3475-3481.

Photo 108389214 © Ivan Shidlovski | Dreamstime.com
Photo 293671412 © Nataliagh | Dreamstime.com
Photo 198092184 © Skypixel | Dreamstime.com