Donor graft failure (GF) is a life-threatening complication of allogeneic stem cell transplantation (SCT). We have performed this analysis to determine the rate and outcome of GF in the era of modern SCT with better supportive care, novel conditioning regimens and more common use of alternative donors. We retrospectively reviewed data of 491 first allogeneic SCTs from HLA-matched siblings (n=284), matched-unrelated (MUD, n=180) or alternative donors (mismatched-related, haplo-identical and cord-blood, n=27); all performed in a single institution since 1/2001. GF was determined when absolute neutrophil count (ANC) had not reached 0.5 × 10*9/L by day 21 (primary GF) or when ANC decreased irreversibly after engraftment within the first 100 days with no signs of recurrent disease (secondary GF). Chimerism testing was performed with FISH for gender markers and PCR for microsatellite polymorphism in sex-mismatched and sex-matched SCT, respectively. In all, GF was diagnosed in 25 patients (pts), cumulative incidence (CI) 5.2% (95%ci 3.5–7.6), 21 primary, 4 secondary. CI of GF was 2.5%, 6.8% and 23.4% after SCT from siblings, MUD or alternative donors, respectively (p<0.001). CI of GF was not significantly different following myeloablative conditioning (14 pts CI 5.7%) or reduced-intensity conditioning (11 pts, CI 4.6%). Pts with predominantly donor population in chimerism testing were given donor cell boost with no additional conditioning. Pts with a predominant host population or rapidly decreasing donor chimerism on successive tests, suggesting immune rejection, were given autologous back-up cells or a second SCT from a different donor. In all, 10 pts were given boost stem-cells from the original donor, 5 pts received a second allogeneic SCT from a different donor with nonmyeloablative, immunosuppressive conditioning (sibling-1, haplo-3, MUD-1), 8 were given autologous back-up cells. Eighteen pts survived > 1 week after second graft infusion and are evaluable for response. Sixteen evaluable pts engrafted, including 6 of 10 given donor booster (2 non-evaluable), all 5 pts given second SCT from a different donor and 7 of 8 given autologous back-up (1 non-evaluable). Engraftment occurred within a median of 10 days (range, 5–15). The probability and pace of engraftment was not different in the different approaches. Eleven pts (44%) were able to be discharged home. Fourteen pts died; 2 early after diagnosis of GF with no intervention, 5 within one week of second graft infusion and prior to engraftment of ongoing infection or organ toxicity, 2 with no engraftment and 5 died early after engraftment from infection (n=3), organ failure (n=1) and GVHD (n=1). With a median follow-up of 19 months (range, 3–68), 6 are alive and 5 additional pts died (relapse-3, GVHD-1, infection-1). The projected 2-year survival for all pts was 23% (95%ci 5–41). Interestingly, 4 pts given autologous cells had donor cell engraftment, 1 had spontaneous autologous reconstitution within 3 weeks, 1 died of fatal GVHD and 1 of infection, both after 2 months with persistent donor cells and 1 remained with donor cell until she relapsed 2 years later. In conclusion, treatment of GF with a chimerism directed method can salvage a subset of pts with GF. Reserving autologous and/or donor backup cells or alternative donor is advisable in pts at high-risk of GF. The observation of allogeneic recovery after autologous stem cell boost is intriguing and of unknown mechanism.

Author notes

Disclosure: No relevant conflicts of interest to declare.