BACKGROUND: A number of automated devices for pretransfusion testing h
ave recently become available. This study evaluated a fully automated
device based on column agglutination technology (AutoVue System, Ortho
, Raritan, NJ). STUDY DESIGN AND METHODS: Some 6747 tests including fo
rward and reverse ABO group, Rh type and phenotype, antibody screen, a
utocontrol, and crossmatch were performed on random samples from 1069
blood donors, 2063 patients, and 98 newborns and cord blood. Also test
ed were samples from 168 immunized patients and 53 donors expressing w
eak or variant A and D antigens. Test results and technician times req
uired for their performance were compared with those obtained by stand
ard methods (manual column agglutination technology, slide, semiautoma
tic handler). RESULTS: No erroneous conclusions were found in regard t
o the 5028 ABO group and Rh type or phenotype determinations carried o
ut with the device. The device rejected 1.53 percent of tests for samp
le inadequacy. Of the remaining 18 tests with discrepant results found
with the device and not confirmed with the standard methods, 6 gave s
uch results because of mixed-field reactions, 10 gave negative results
with A(2) RBCs in reverse ABO grouping, and 2 gave very weak positive
reactions in antibody screening and crossmatching. In the samples fro
m immunized patients, the device missed one weak anti-K, whereas stand
ard methods missed five weak antibodies. In addition, 48, 34, and 31 o
f the 53 weak or variant antigens were detected by the device, the sli
de method, and the semiautomated handler, respectively. Technician tim
e with the standard methods was 1.6 to 7 times higher than that with t
he device. CONCLUSION: The technical performance of the device compare
d favorably with that of standard methods, with a number of advantages
, including in particular the saving of technician time. Sample inadeq
uacy was the most common cause of discrepancy, which suggests that sta
ndardization of sample collection can further improve the performance
of the device.