IntroductionIt has been previously recognised that unrelated haematopoietic stem cell donors from certain ethnic minority groups are less likely to be available when called to give a sample for confirmatory (verification) typing (CT), prior to final donor selection. However, a direct impact of such donor attrition on donor selection has not, to date, been shown for patients of any ethnicity. MethodsWe analysed the outcome of CT requests for 303 patients managed by the Anthony Nolan Graft Identification and Advisory Service (GIAS), with the aim of assessing the impact of donor attrition on the speed of donor provision and donor suitability at CT and at transplant. Patient ethnicity was categorised as white Northern European (WNE) or non-white Northern European (non-WNE). A donor was categorised as unavailable if they were unable to return a sample for CT within the timeframe required by the transplant centre, for whatever reason. Requests cancelled by the transplant centre within this timeframe were not categorised as donor attrition. Results235 patients (77.6%) were of white Northern European (WNE) descent, and 68 (22.4%) of non-WNE descent. 43 patients (13.0%) had a high, 141 (42.5%) a medium, 127 (38.3%) a low, and 21 (6.3%) a very low probability of search success (identifying a 10/10 matched donor), based on DRB1 allele and DRB1-DQB1 haplotype frequencies. Donor attrition was more common in those donors requested for CT for non-WNE patients (38.4% vs 24.6%, p<0.001), and more non-WNE patients were affected by at least two incidences of attrition (40.2% vs 29.0%, p=0.059). Patients experiencing two or more incidences of donor attrition had a longer time from first CT request to identifying the donor eventually selected for transplant (median 26 days vs 34.5 days, p<0.001). In addition, donor attrition was associated with fewer patients with at least one 10/10 donor identified at CT (46.5% vs 65.8%, p=0.001), fewer patients with a CMV matched donor at transplant (67.6% vs 80.4%, p=0.044) and fewer patients with a donor aged 30 or less at transplant (44.9% vs 62.8%, p=0.014). When analysing WNE patients alone, similar findings were made. However, the impact of attrition on search outcomes was less for non-WNE patients. Of note, there was only a trend towards a difference in time from first CT request to identifying the donor eventually selected for transplant (29.5 vs 38.5 days, p=0.072), and no impact on number of 10/10 donors identified at CT (22.6% vs 27%, p=0.673). In addition, no impact on CMV status or donor age at transplant was noted for non-WNE patients. Since patients with less common HLA phenotypes are likely to need a greater number of CT requests (and therefore are more exposed to donor attrition), a multivariate analysis was carried out including the probability of search success. For WNE patients, 2 or more episodes of donor attrition remained significantly associated with fewer patients with a 10/10 match at CT (p=0.047), more CMV mismatches at transplant (p=0.043), and fewer donors aged 30 or under (p=0.03). ConclusionWe have shown a significant impact of donor attrition on both search times and the availability of well-matched donors. Donor attrition affects not only the provision of HLA-matched donors, but also those matched for secondary characteristics such as CMV and age. However, despite a higher proportion of non-WNE patients experiencing donor attrition, the overall impact of attrition was less when compared to WNE patients. This may be because non-WNE are already sufficiently disadvantaged by the lack of available HLA-matched donors on international registries that the impact of donor attrition is minor in comparison.Since these variables are strongly associated with survival, donor attrition must be addressed at the donor registry level in order to improve patient outcome following transplant. Disclosures:No relevant conflicts of interest to declare.