A comparison of two automated methods for the detection and identification of red blood cell alloantibodies

The aim of this study was to compare the routine use of two automated systems (OrthoAutoVue Innova, microcolumn, and Immucor Galileo, solid phase) for the screening and identification of irregular red blood cell alloantibodies in samples, analysed in our Transfusion Service during 6 months of normal activity. The study focused particularly on an evaluation of the repeatability of the screening tests, the identification of antibody specificities and the identification of antibodies in samples showing discordant results.

Overall 2,229 samples from potential blood donors (A), multiply transfused patients with blood disorders (DH), potential transfusion recipients (TS), and external cases (E) were studied. The protocols were carried out according to the manufacturers recommendations.

The screening tests detected 78 samples that were positive with both systems, while 18 were positive only with Immucor and 11 only with Ortho (thus, overall, Immucor detected 96 positive samples and Ortho 89 positive samples). The use of the respective identification panels enabled us to identify the antibodies in 65 samples with Immucor and in 61 samples with the Ortho system; 74 antibodies were identified with Immucor (55 with a single specificity and 19 with mixed specificities) and 68 antibodies with Ortho (51 and 17, respectively). In the remaining cases (31 samples for Immucor and 28 for Ortho), the antibody specificity was not identified. The two systems were found to be essentially similar. The Immucor system revealed a greater number of antibodies, mainly because of its greater sensitivity at detecting anti-D antibodies.

Both systems showed a repeatability of over 85%, demonstrating that automation of immunohaematological tests is advantageous. The specificity of the antibody was identified in 68% of the samples. Furthermore, using the two systems led to the identification of ten new antibodies (6 anti-D, 2 anti-E, 1 anti Le(a), and 1 anti-Vel), which would not have been detected had only one of the two methods been used.