[Comparison of the frequencies and distributions of unexpected antibodies based on different calculation criteria].

Background: The frequencies and distributions of unexpected antibodies have been reported using two different criteria, based on either number of persons tested or number of tests performed. But there has been no study that compared the results of analyses based on these two different criteria using the same data set.

Methods: Unexpected antibody tests performed in a University Hospital during recent 6 yr (January 2002-December 2007) were retrospectively analyzed: 76,985 tests (59,503 persons) for screening and 875 tests (749 persons) for identification. Data were analyzed using two different criteria, based on 'persons tested' and 'tests performed'. Antibodies had been screened and identified using LISS/Coombs gel cards with DiaMed-ID system (DiaMed AG, Switzerland).

Results: Frequencies of unexpected antibodies based on 'persons tested' and 'tests performed' were 1.32% and 1.34%, respectively (P=0.88). For frequently detected as well as rarely detected antibodies, there were no significant differences in the frequencies based on two different criteria. However, for rarely detected antibodies (anti-Xg(a) and Anti-E & D), the frequencies based on 'tests performed' were higher than those based on 'persons tested', affecting a change in the order of frequencies of antibodies detected.

Conclusions: As there were no significant differences in the frequencies of unexpected antibodies calculated using two different criteria, both criteria can be used together for the patient population in our hospital. However, two criteria should be compared to validate the results for other populations.