Sample Identification Errors Observed Within A Study for Molecular Cord Blood Hla Typing

Background: We studied the practicability of low-resolution HLA class I molecular or serological typing in addition to medium-resolution class II DNA typing for maternal and cord blood (CB) samples. Material and Methods: 103 randomly selected sample pairs were evaluated. Results: In one of the CB specimens a third antigen at each HLA locus occurred due to a contamination of the CB by maternal cells. In 3 pairs (2.9%) the HLA antigens of the CB sample did not match the maternal antigens as a result of sample mix-up. Discussion: Routine amplification refractory mutation system (ARMS) class I typing for all CB specimens is not yet a feasible method, but it should be performed when the result of serological typing is ambiguous. Additionally, we conclude that it is both important and practical to type the maternal HLA antigens from a frozen back up sample collected at birth when the CB is considered for clinical use, and to compare the typing result with the previously typed CB transplant in order to expose any sample identification errors.