Comparison of a commercial immunochromatographic strip crossmatch kit and standard laboratory crossmatch methods for blood transfusion compatibility in dogs

Objective To evaluate agreement between 2 standard laboratory (SL) methods and an immunochromatographic strip (ICS) method to crossmatch dogs receiving RBC transfusions. A second objective was to evaluate uninterpretable SL crossmatch results as compared to ICS in the presence of autoagglutination. Design Prospective observational study (September 2018 to October 2019). Setting University teaching hospital. Animals Forty anemic dogs receiving RBC transfusions. Interventions None. Measurement and Main Results All dogs received DEA 1-negative packed RBCs. Three crossmatch methods were evaluated against the same unit transfused to each dog: SL method performed at institutional laboratory (SL-I), SL method sent to a commercial laboratory (SL-C), and a commercially available point-of-care ICS method. Major and minor crossmatches were incompatible for 2.5%/7.5% of ICS tests, 82.5%/52.5% of SL-I tests, and 52.5%/27.5% of SL-C tests. Agreement between ICS and SL-C major (kappa = 0.05) and minor (kappa = 0.02) crossmatches and between ICS and SL-I major (kappa = 0.009) and minor (kappa = 0.03) crossmatches was slight. Agreement between SL-C and SL-I major (kappa = -0.06) and minor (kappa = -0.12) crossmatches was poor. Results of major and minor crossmatches were uninterpretable due to autoagglutination in 38%/38% for SL-I and 29%/18% for SL-C crossmatches. ICS method was interpretable for 93% (major) and 98% (minor) crossmatches. After exclusion of uninterpretable SL pairings, agreement still remained poor to slight between all tests. Only 1 of 40 dogs (2.5%; 95% confidence interval: <1.0%-13.2%) had an immediate immunological transfusion reaction. Conclusions Lack of agreement between all methodologies was noted. The high level of incompatibility predicted by SL methods despite lack of clinically relevant reactions suggests a high false incompatibility rate as compared to the ICS test. ICS testing was also able to give results more frequently in the face of autoagglutination. Further work is needed to investigate the ICS method's ability to predict clinically significant transfusion reactions.