Determining Late Engraftment Following Single Cord, Unrelated Transplantation: An Analysis of the Eurocord Registry

The most prominent clinical problem in CB transplant is a slower engraftment kinetic, as compared to the conventional SC sources. Engraftment failure is usually defined in CB transplantation as a lack of PMN recovery 60 days after transplant, whilst a definition of late engraftment is currently lacking. A late engraftment is a major cause of transplant-related mortality (TRM), due to the prolonged exposure of the recipient to the infective and hemorrhagic risk. After the HSCT, the probability of engraftment at each time interval tends to increase up to a maximum and then gradually decreases, which is typically represented by a sinusoid curve. The decrease of engraftment probability after transplant results in a rapidly increasing risk of TRM; therefore such a turning point should be considered the beginning of a risk phase for late/no engraftment. Therefore, it is important to find the time point after UCBT in which the probability of engraftment will decrease in order to help taking a decision for rescue with a second transplant. We analyzed the clinical expectations beyond this time in a homogenous population of CB recipients.

We investigated the engraftment kinetic in a population of 1215 patients who received a single, unrelated CB transplant for Acute Leukemia (AL) in Complete Remission (CR) following a Myeloablative Conditioning Regimen (MAC). All patients were transplanted in EBMT Centers and reported to the Eurocord Registry from 1994 to 2010. Ratio Lymphoid/Myeloid Leukemias was 769/445, reflecting a major proportion of pediatric patients over adults (857/357). Patients were transplanted in first (43.4%), second (46,6%), or third or subsequent remission (10%), respectively. Median (range) age at transplant was 9.5 (0.3-63) years. Median weight (Kg) at transplant was 33 (5-112). Out of 1089 patients evaluable for HLA-matching, 601 (55.2%) were mismatched for 0–1 loci, 448 (41.1%) for 2 loci and 40 (3.6%) for more than 2 loci. Fifty percent of the patients had a TBI-based myeloablative regimen. Data on TNC counts at freezing of transplanted CBU were available in 963 cases: median and range were 5 (1.1-41.83)x10⁷/Kg.

The median FU was 30 months (1-174). At 24 months overall survival was 49±2%, TRM was 32±2%. Median time of engraftment was 24 days (10-133) with a cumulative incidence of 86±1% at day 60. Analyzing the cumulative curve of engraftment, we considered the engraftment probability within intervals of five days after the transplant; in fact the highest probability of engraftment was at day 25 and dropped of 50% at day 42. Among 167 patients (13.7%) who did not engraft at this time, 63 patients (38%) experienced a late engraftment with a median time of 47 days (43-131) after transplant. The cumulative incidence of engraftment at 120 days was 37% and 38% at day 180 without any further increasing later on. Out of the 104 patients who never engrafted, 74 died and major causes of death were bacterial (17%), viral (10%) and fungal (9%) infections, respectively, whilst 30 patients are alive at the last follow up. Information of graft failure treatment was available for 84 patients. Twenty eight did not receive any treatment (25 died at a median time of 80 days form UCBT), 24 had an autologous back up and 32 underwent a second allogeneic HSCT (14 second UCBT, 9 Haplo PBSC and 9 unrelated BMT). Of those 32, 17 patients engrafted, 5 relapsed; 24 died, 8 are alive at last follow up.

The maximum probability of engraftment after UCBT for patients with AL in remission is at day 25 and halves at day 42, thus suggesting that a clinical decision should be made within this period. In particular, rescue actions, such as infusion of another graft, either allogeneic of autologous, should be considered. Such a model can be applied to different subsets of patients and is particularly useful in transplant at high risk of late engraftment such as UCBT.

No relevant conflicts of interest to declare.