HLA Matching in Cord Blood Transplantion: Clinical Outcome and Implications for Cord Blood Unit Selection and Inventory Size and Ethnic Composition.

Background: Engraftment and survival in cord blood transplantation is affected by total nucleated cell (TNC) dose and HLA match grade. Thus, patients need assess to more cord blood units (CBUs) to provide for better matches and higher cell dose.

Purpose: To examine the effect of HLA match grade and direction of mistmatch on outcome of cord blood transplantation with the aim to improve match algorithms and cord blood unit selection strategies. To apply the match algorithms in an empirical analysis of the chance of finding a CBU match by match grade, taking into account inventory size and donor ethnic background.

Materials and Methods: Engraftment, GvHD, relapse, TRM and overall survival was evaluated in 1511 recipients of single CBU from the NYBC National Cord Blood Program transplanted through June 2005 (data available on 91% of eligible patients). The empirical chance of finding a full or 1 antigen mismatch (HLA-A, B at intermediate level and DRB1 at high resolution ) CBU was assessed for 14,378 patients on an inventory 25,917 CBUs from ethnically diverse donors (16.5% African-American, 17.0% Hispanic, 6.6% Asian, 50.4% Caucasian and 9.6% mixed).

Results: HLA match grade correlated with engraftment, GvHD, relapse, TRM and overall survival, independent of the effect of TNC dose and other contributing factors. TNC dose/kg patient body weight (range 0.7 to >10.0 x 10^7/kg) did not affect outcome in fully matched CBU grafts. A two-fold increase in TNC dose, on average, was required to overcome differences in TRM and survival for 2 antigen mismatched grafts compared to 1 mismatch. Unidirectional mismatches in GvHD direction only (CBU homozygous at the mismatched loci) provided engraftment, TRM and survival rates that were the same as fully matched grafts, also with no apparent relationship to TNC dose. Patients with leukemia given CBU grafts with a GvHD only ismatch had a relapse rate that was the same at that of fully matched CBUs, whereas those with rejection only mismatches had a significantly higher relpase rate. In the empirical analysis of the chance of finding a match, including mismatch direction, 10.6% of patients could find a full match in an ethnically diverse CBU inventory. Including CBUs with 1 or 2 antigen mismatches only in the GvHD direction more than doubled the chance of finding a "good" match. The chance of finding a match or a CBU with 1 antigen mismatch correlated with patient ethnicity (patients had the best chance of finding a good match within their own ethnic group). African-American patients were the least likely of any group to find a suitable CBU (more than 5-fold less likely than Caucasions to find a full match), in part, because of their smaller numbers in the inventory and, in part, because of their greater HLA diversity. Increasing inventory size increased the chance of finding a full or 1 antigen matched CBU for all ethnic groups, but the improvement diminished progressively.

Conclusions: HLA match level, including direction of mismatches, was associated with success of CBU transplantation. As with marrow donor registries, improvement in the probability of matching requires a progressively larger inventory size and depends on the donor ethnic diversity. Quantitatively larger number of donors are required for ethnic minority patients. Thus, the US National Cord Blood Inventory (planned for 150,000 new, high quality CBUs) should be enriched with ethnical diverse donors in order to meet the needs of ethnic minority patients who lack an HLA identical sibling donor.