A multicenter study on the performance of a fully automated, walk-away high-throughput analyzer for pretransfusion testing in the US population.

BACKGROUND: Moving to automation is a major focus of transfusion centers. Erytra (Grifols) is a walk-away analyzer with high-performance and -throughput capacity for pretransfusion testing. Efficiency and performance of Erytra with its cards and reagents were evaluated in comparison to Food and Drug Administration (FDA)-approved reference methods. STUDY DESIGN AND METHODS: A total of 5279 blood samples (46% patients; 54% donors) were obtained from US blood establishment facilities. Samples were analyzed with Erytra and results were compared with the routine FDA-licensed automated platforms used by the clinical study sites. A total of 25,217 tests were performed (15,322 ABO/D/reverse typing; 4916 Rh phenotypes, 669 K typing, 838 antibody screens, 759 antibody identifications, 250 cross-matches, 244 ABO compatibilities by immediate-spin cross-match, and 219 direct antiglobulin tests [DATs]). RESULTS: Global agreement between Erytra and the comparison platforms was 99.66%, with 99.82% positive percent agreement (95% lower confidence bound [LCB], 99.75%) and 99.50% negative percent agreement (95% LCB, 99.37%). There were 85 discrepancies (0.34%), including cross-matches (n = 13), antibody screens (n = 10), antibody identifications (n = 21), and DATs (n = 5), whereas an excellent concordance was obtained in blood grouping determinations (ABO/D/C/E/c/e/K, 0.04%-0.22% discrepancies). Analysis of the discrepancies showed that Erytra provided the correct result in 51 of them (60%), with only five false negatives (one O patient transplanted with A, one mixed-field reaction in a very weak D, one anti-Vel, two A2rr). Erytra results were 100% reproducible in a series of 3760 repetition tests. CONCLUSION: Grifols' Erytra analyzer showed reliable efficacy compared with equivalent FDA-licensed reagents and FDA-cleared instruments.

Measurements of human herpesvirus 8 viral load in blood before and after leukoreduction filtration.

BACKGROUND: Human herpesvirus 8 (HHV-8) is likely transmitted through blood transfusion in high-prevalence areas. The efficacy of leukoreduction filtration for reducing HHV-8 in blood has not been reported. STUDY DESIGN AND METHODS: Blood was drawn from 45 human immunodeficiency virus-positive men either with Kaposi's sarcoma (KS; n=21) or without KS (n=24) and subject to leukoreduction filtration. HHV-8 viral load was measured in plasma and in blood before and after filtration. RESULTS: Twelve subjects, all with KS, had detectable HHV-8 viremia before filtration with viral loads of 10(2) to 10(5) copies/mL (mean, 3 × 10(4) copies/mL). After filtration, seven of 12 subjects no longer had detectable HHV-8 in their blood, and five of 12 subjects had detectable HHV-8 that was 90% reduced on average from prefiltration levels. The presence of HHV-8 in the blood after filtration was strongly associated with prefiltration viral loads greater than 1000 copies/mL and the presence of cell-free virus in plasma. None of the subjects without KS had detectable levels of HHV-8 virus in blood before or after filtration. CONCLUSION: Cell-associated HHV-8 appeared to be effectively removed by leukoreduction filtration. Cell-free HHV-8 was present in 42% of subjects as 1% to 20% of the total virus which was not removed by filtration.

Improved method for fluorescence cytometric immunohematology testing.

BACKGROUND: A method for accurate immunohematology testing by fluorescence cytometry (FC) was previously described. Nevertheless, the use of vacuum filtration to wash RBCs and a standard-flow cytometer for data acquisition hindered efforts to incorporate this method into an automated platform. STUDY DESIGN AND METHODS: A modified procedure was developed that used low-speed centrifugation of 96-well filter plates for RBC staining. Small-footprint benchtop capillary cytometers (PCA and PCA-96, Guava Technologies, Inc.) were used for data acquisition. Authentic clinical samples from hospitalized patients were tested for ABO group and the presence of D antigen (n = 749) as well as for the presence of RBC alloantibodies (n = 428). Challenging samples with mixed-field reactions and weak antibodies were included. Results were compared to those obtained by column agglutination technology (CAT), and discrepancies were resolved by standard tube methods. Detailed investigations of FC sensitivity and reproducibility were also performed. RESULTS: The modified FC method with the PCA determined the correct ABO group and D type for 98.7 percent of 520 samples, compared to 98.8 percent for CAT (p > 0.05). No-type-determined (NTD) rates were 1.2 percent for both methods. In testing for unexpected alloantibodies, FC determined the correct result for 98.6 percent of 215 samples, compared to 96.3 percent for CAT (p > 0.05). When samples were automatically acquired in the 96-well plate format with the PCA-96, 98.7 percent of 229 samples had correct ABO group and D type determined by FC, compared to 97.4 percent for CAT (p > 0.05). NTD rates were 0.9 and 2.6 percent, respectively. Antibody screens were accurate for 99.1 percent of 213 samples with the PCA-96, compared to 99.5 percent for CAT (p > 0.05). Further investigations demonstrated that FC with the PCA-96 was better than CAT at detecting weak anti-A (p < 0.0001) and alloantibodies. CONCLUSIONS: An improved method for FC immunohematology testing has been described. This assay was comparable in accuracy to standard CAT techniques, but had better sensitivity for detecting weak antibodies and was superior in detecting mixed-field reactions (p < 0.005). The FC method demonstrated excellent reproducibility. The compatibility of this assay with the PCA-96 capillary cytometer with plate-handling capabilities should simplify development of a completely automated platform.

An automatable format for accurate immunohematology testing by flow cytometry.

BACKGROUND: Current immunohematology testing methods have limitations including cost, throughput, and adaptability to automation. Furthermore, current automated and semiautomated workstations cannot accommodate many other tests relevant to blood transfusion. STUDY DESIGN AND METHODS: Authentic clinical samples from hospitalized patients were tested for ABO group, D type, and presence of RBC alloantibodies by column agglutination technology (CAT), standard tube methods, and a recently developed flow cytometry (FC) technique. Included were challenging samples with rouleaux, autoantibodies, mixed-field reactions, and weak antibodies. Antibody staining of RBCs for FC was initially performed in test tubes and subsequently in microtiter filter plates interfaced with a vacuum manifold. RESULTS: When antibody staining was performed in tubes, FC testing determined the correct ABO group and D type for 99.1 percent of 222 clinical samples, as compared to accuracies of 91.9 percent for CAT and 95.0 percent for standard tube testing. FC testing also detected 99.5 percent of clinically relevant RBC alloantibodies in 239 patient samples, as compared to 98.9 percent for CAT and 94.7 percent for LISS-IAT. Using the FC filter plate technique, 104 of 109 samples (95.4%) were correctly typed for ABO and D (the remaining five samples were read as "no type determined" due to RBC and serum testing discrepancies), and RBC alloantibodies of the IgG and IgM classes were correctly identified in 98.3 percent of samples. CONCLUSIONS: Optimized FC testing methods that are comparable in accuracy to standard CAT and tube methods are described. When used with filter plates, this methodology should allow rapid and cost-effective immunohematology testing of both patient and donor samples in an automated workstation format. The same workstation should support automation of other pretransfusion assays that can be analyzed by FC.