Frequency and Risk Factors of Cord Graft Failure (CGF) Following Reduced Intensity Conditioning Haplo-Cord Hematopoietic Stem Cell Transplantation

PURPOSE: Delayed engraftment and cord graft failure (CGF) are serious and often fatal complications after unrelated cord blood (UCB) hematopoietic cell transplantation (HCT), precluding use of low cell dose UCB. The haplo-cord HCT approach allows the use of a lower dose single UCB unit by co-infusion of a CD34+ selected haploidentical graft. Although haplo-cord HCT aims to achieve durable UCB hematopoiesis, the haplo graft provides early temporary engraftment. We describe the frequency, complications and risk factors of CGF after haplo-cord HCT after reduced-intensity conditioning (RIC).

PATIENTS AND METHODS: Adult hematologic malignancy patients from the University of Chicago or the Weil Cornell medical centers who underwent haplo-cord HCT between 2007 and 2013 were included. Conditioning consisted of fludarabine, melphalan, and rATG (and TBI 400 cGY if high CNS relapse risk), followed by infusion of a CD34+ selected G-CSF mobilized haploidentical graft and the best HLA matched single UCB unit of at least 0.5 or 1.0 x 10^7 total nucleated cells (TNC)/kg, depending on the protocol. CGF was defined as <5% cord blood chimerism by Day 60 in the unfractionated or CD3 compartments irrespective of neutrophil or platelet counts. Death before Day 60 excluded patients from the primary outcome of CGF at Day 60. Univariate analyses were performed to identify potential risk factors for CGF: Fisher's Exact test for dichotomous and logistic regression for continuous predictor variables.

RESULTS: 107 patients were evaluated. Chimerism data were not available on two, and 11 (10.3%) died before Day 60, leaving 94 evaluable patients for CGF. Diseases indications were: AML (51%), ALL, (12%), MDS (11%), and other (25%). Median age of patients was 50 years (range 18-73) and many had active disease at HCT (47%). The mean UCB collected dose was 2.1x10^7 TNC/kg (range 0.77-8.3x10^7 TNC/kg) and HLA cord match was 4/6 in 24% and 5/6 or 6/6 in 73%. Few patients had UCB doses below 1x10^7 TNC/kg (N=5).

CGF occurred in 14 of 94 (15%) evaluable patients. Of these, 7 died within 1 year of transplant date. The causes of deaths were: poor graft function (N=2), infection (N=2), relapse or progressive disease (N=2) and unknown (N=1). The other 7 remain alive (range: 7 months to 5.5 years) with haplo-derived or mixed haplo-recipient hematopoiesis. Four are in remission and three have relapsed disease. Median survival for the CGF group was 12.7 months.

In univariate analyses, no UCB factor, including cell doses, major ABO mismatch, donor specific antibodies, CMV status, and HLA-match, was associated with CGF. (Table) However, higher haploidentical TNC and CD34+ doses were associated with greater risk of CGF.

CONCLUSION: Approximately 15% of patients experienced CGF after haplo-cord HCT. Most have died or relapsed, but some have had relatively long-term survival due to sustained haploidentical hematopoiesis. Avoidance of high haploidentical cell doses may reduce risk of CGF. We were unable to identify other determinants of CGF. In ongoing studies, we have limited the haplo graft to <5 x10^6 CD34+/kg and are testing the use of lower UCB cell doses when large units can not be identified. Additional follow-up is needed to determine if sustained or greater cord chimerism improves long-term outcomes compared to haplo chimerism after RIC haplo-cord SCT.