The aim was to determine whether outcome of unrelated donor transplantation for severe aplastic anemia has improved in recent years and whether this is due to patient selection or better transplant technology. We analyzed 498 patients transplanted during 1990-2005. By running univariate regression models dichotomizing year of transplantation we defined 1998 as the year of the most significant change in survival. Five-year survival increased from 32+/-8% before 1998 to 57+/-8% after 1998 (P<0.0001). When comparing the cohort before (n=149) and after 1998 (n=349), there were no differences except for older age, and more frequent use of PBSCs, after 1998. High-resolution HLA typing data were unavailable. After 1998, there was less graft failure (11 vs 26%, P<0.0001), less acute GvHD (cumulative incidence 28 vs 37%, P=0.02) and less chronic GvHD (22 vs 38%, P=0.004). In multivariate analyses adjusting for differences in age, HLA-mismatch, performance score and time to transplantation, there was no change in the year of transplant effect (relative risk of death in transplants after 1998: 0.44 (95% confidence interval 0.33-0.59)). There is no evidence for patient selection to explain significantly improved survival in patients transplanted after 1998. We speculate that this is due to better donor matching.