Quality control of a national bone marrow donor registry: results of a pilot study and proposal for a standardized approach.

Unrelated hematopoietic stem cell transplantation (HSCT) is a recognized therapy for hematological diseases and over 8 million HLA-typed donors are ready to donate. Increased international exchanges and rapid requests through the Bone Marrow Donor Worldwide (BMDW) ask for standardized quality assurance. Since no such standards have been established to date, we tested a pilot program in order to evaluate donor availability and quality of HLA typing of the Swiss Registry. The 18500 donors of the registry have been analyzed by serology for HLA-AB and by molecular typing for HLA-DR. Through three successive annual quality control (QC) exercises, a total of 114 donor requests were sent to 13 blood transfusion centers responsible for donor recruitment asking for a blood sample. Donors were randomly selected according to recruitment periods (1988-1993; 1994-1997; 1998-2000), and to homozygosity for HLA-A and/or -B antigens. An additional 80 frozen blood samples from the repository corresponding to the three periods (n=26) and to the 2001 period (n=54) were also included in the HLA study. HLA-AB typings were done by polymerase chain reaction-sequence specific primers (PCR-SSP) and all discrepancies were retyped. The results showed that 79 samples provided by 69.3% of the requested donors were received within 14 days, and 19 samples (16.7%) were received in >14 days. Altogether, an 86% rate of donor availability was observed, independent of the recruitment period. Among the requested donors, 16 (14%) were not available: for medical reasons (two), for personal reasons (eight), for loss (one), and for an unknown reason (five). The HLA-A/B DNA typing results of 166 homozygous and 12 heterozygous blood samples showed that 437/439 (99.5%) of the assigned A/B antigens were correct. However in 36/178 donors (20.2%) an HLA-A or -B antigen had been missed (34 donors) or misassigned (two donors) by serology, with a decreasing discrepancy rate of 30% (1988-1993) to 18.5% in 2001. Assuming that HLA-A or -B homozygotes are found in 10-15% of the donors and that correct assignments have been observed in nearly 100% of the donors, an overall error rate of 4-5% would be expected for the national registry HLA-AB typing. These data show that standardized quality control for donor availability and HLA typing is feasible, and we propose that this model could be applied to the registries participating in bone marrow donor worldwide.