Double cord blood transplantation in patients with high risk bone marrow failure syndromes


Professor G. Socié, Service d’Hématologie-Greffe de Moelle, Hôpital Saint Louis, APHP, Paris University 7, 1 avenue Claude Vellefaux, 75010 Paris, France. E-mail:


Patients with bone marrow failure syndromes (BMFS) who reject a first allogeneic transplant or fail immunosuppressive therapy (IST) have an especially grim prognosis. We report 14 patients (eight adults, six children) transplanted with double cord blood transplantation (dUCBT) for BMFS. Neutrophil recovery was observed in eight patients, with full donor chimerism of one unit, and acute GVHD in 10. With a median follow-up of 23 months, the estimated 2 years overall survival was 80 ± 17% and 33 ± 16% for patients with acquired and inherited BMFS, respectively. Transplantation of two partially HLA-matched UCB thus enables salvage treatment of high-risk patients with BMFS.

Acquired bone marrow failure syndromes (BMFS), such as severe aplastic anaemia (SAA), or inherited BMFS, such as Fanconi anaemia (FA), are rare. The outcome of these patients has greatly improved in recent years but is still poor for patients who lack a sibling donor and who failed or relapsed after immunosuppressive therapy (IST) or who have suffered a graft failure of previous allogeneic hematopoietic stem cell transplantation (HSCT). The use of alternative donors has been associated with poor long-term survival (Bacigalupo et al, 2000; Gluckman et al, 2007). Recently, transplantation from unrelated donors has improved significantly due to better human leucocyte antigen (HLA) donor matching, optimization of the conditioning regimen, and improvement in supportive measures (Deeg et al, 2006; Locasciulli et al, 2007). The prognosis remains extremely poor for patients with inherited BMFS or for those with acquired BMF who fail IST and who lack a suitable alternative donor and for those with previous rejection. Umbilical cord blood transplantation (UCBT) has extended the availability of HSCT in the absence of a suitable donor and can be used in urgent situations, such as graft failure (Fernandes et al, 2007). However, the cell dose obtained from a UCB unit limits its use in adolescent and adults recipients with BMFS, in whom a valid stem cell dose is crucial to achieve engraftment (Rocha & Gluckman, 2006). To overcome the cell dose barrier, the use of double UCBT (dUCBT) has been successful in achieving a high engraftment rate in high-risk haematological disease (Brunstein et al, 2007). To date dUCBT has been reported in only few patients with BMFS (Mao et al, 2004). In order to enhance engraftment and to try avoiding rejection, we conducted a study to evaluate dUCBT in 14 patients diagnosed with high risk BMFS.

Patients, materials and methods

Fourteen consecutive patients (six male and eight female) with a median age of 16 years (range 6–31 years), diagnosed with BMFS [eight FA, four SAA, one dyskeratosis congenital (DC), one Paroxysmal Nocturnal Hemoglobinuria (PNH)] underwent a dUCBT from 2004 to 2007. All patient samples and data were collected after obtaining informed consent in accordance with the Declaration of Helsinki, after approval by the Institutional Review Board of the hospital Saint Louis, Paris. The median weight at transplantation was 45 kg (range 17–72 kg). Conditioning therapy consisted of a fludarabine (100 mg/m2 total dose)-based regimen for all patients, with total body irradiation (TBI; 2 Gy) for four patients. Antithymocyte globulin (ATG) was added in 11 cases (79%). UCB unit that was 4–6/6 HLA-A, B (antigen typing) and DRB1 (allele typing) matched to the recipient was selected (UCB number 1). UCB number 2 had to be both 4–6/6 HLA-A, B and DRB1 matched to the recipient and to UCB number 1. The two UCB units were mismatched with the recipient for one (n = 3), two (n = 9) or three (n = 1) HLA loci. One patient received two complete HLA-matched units. The UCB units were 6/6 HLA matched to each other in three patients, 5/6 in three patients, and 4/6 in eight patients.

The median numbers of infused nucleated and CD34+ cells were 4·8 × 107/kg (range 1·8–9·7) and 2·9 × 105/kg (range 0·5–7·46) for the combined 2 UCB units, respectively. The median infused cell doses for each individual UCB unit were 2·2 × 107 nucleated cells/kg (range 0·9–8·40) and 1 × 105 CD34+ cells/kg (range 0·3–6·48). Cyclosporine A (CSA) alone (3 mg/kg daily) or combined with mycophenolate mophetil (MMF) (15 mg/kg twice daily) was used as acute graft-versus-host disease (GVHD) prophylaxis in nine and five cases, respectively. Steroids (1 mg/kg daily) were given from day 7 to day 14 and stopped in the absence of active aGVHD at day 28. Details concerning transplantation procedures and clinical care in our unit have been described elsewhere (Bittencourt et al, 2002). Time to neutrophil engraftment was defined as the first of three consecutive days with an absolute count >0·5 × 109 cells/l. Full donor chimerism was defined as marrow reconstitution of greater than 90% of donor cells. Event times were measured from day 0 of UCB infusion to the date of death or last contact. Kaplan-Meier method was used to calculate overall survival (OS). Cumulative incidence using death as a competing event was performed to estimate incidence of neutrophil recovery and aGVHD. The retrospective analysis was performed as of April 2008. Statistical analyses were performed with SPSS (Inc., Chicago) and S-Plus (MathSoft) software packages.

Results and discussion

Details on patients and graft characteristics are shown in Table I. All five patients with acquired SAA had failed two prior courses of IST, and two had rejected a previous allogeneic graft. Four out of nine patients with inherited BMFS were transplanted because of a previous rejection. Two patients with FA underwent transplantation whist in leukemic transformation. The median time from diagnosis to transplantation was prolonged (median 31 months, range 5–240 months), showing that dUCBT was the last option of treatment for those patients. Therefore, all patients were heavily transfused prior to transplantation (seven patients had received more than 20 red blood cell units and six patients more than 20 platelet transfusions). They were all at increased risk of graft rejection, because of the potential alloimmunization induced by multiple transfusion exposure and the long median disease duration (Ades et al, 2004). As institutional practice, all pretransplant recipients’ serum samples were tested for the presence of HLA class I and class II antibodies. No HLA-specific alloantibodies were detected against HLA antigens of the UCB units. Furthermore, at the time of dUCBT the median neutrophil and platelet count was 0·3 × 109/l (range 0·0–1·1) and 21 × 109/l (range 3–93), respectively. Six patients (43%) received a dUCBT as rescue of a previously rejected transplant (three FA, two SAA and one DC). Among these latter patients, one has been already reported by our group (Fernandes et al, 2007). The median time to achieve neutrophil engraftment was 28 d (range 14–42 d). The cumulative incidence (CI) of neutrophil recovery at day +60 was 57 ± 14% (n = 8). The median time to platelet recovery (>20 × 109cells/l) was 105 d (range 36–180 d). The CI for platelet recovery at day +180 was 21 ± 14%. Graft failure was observed in six patients, three of whom were transplanted for a first non-engraftment. All patients received sufficient nucleated cells with a good CD34+ cell viability (data not shown), indicating that graft rejection was probably due to an immune mechanism. Moreover, donor-recipient HLA matching may also explain failure of engraftment in these high-risk patients, as almost all patients (10 of 14) received two HLA mismatched UCB units. HLA disparities (0, 1 or 2 vs. 3 and 4 of 6) is an important factor associated with engraftment after single unit UCBT for patients with malignant disorders (Wagner et al, 2002). Cell dose can overcome the HLA barrier in single UCB transplantation when the number of HLA disparities is not more than 3. However, in non-malignant disorders including FA, the numbers of HLA disparities, including two or more, are associated with lower probability of neutrophil recovery and decreased survival (Gluckman et al, 2007).

Table I.   Patients characteristics. Thumbnail image of

At day-100 post-dUCBT, donor-recipient chimerism analysis was evaluable in nine patients (72%). Seven patients showed a full donor chimerism with a single UCB unit engrafted; in one patient a combination of both UBC units was detected (89% first UCB infused, 11% second UCB unit infused), and one patient had an autologous recovery. In six out of eight patients, the predominant UCB unit was the first infused, as described by others (Ballen et al, 2007). After day 100, all evaluable patients (n = 7) were complete donor chimera. No patients had experienced late rejection at the last assessment.

The CI of grade II-IV aGvHD was 71 ± 13% (n = 10). It was of grade II in six cases (43%), III in three cases (21%) and IV in one case (7%) (Glucksberg et al, 1974). This high aGvHD incidence might be related to patients’ characteristics, since all the FA group patients (eight out of 14) developed aGVHD. FA is already known to be associated with severe acute GvHD after transplantation from matched related donors (Guardiola et al, 2004). The impaired DNA repair process, typical of FA, allows tissues to become a target of the alloreactive aggression and impacts both on the occurrence and severity of aGVHD. Also in recipients of dUCBT, the risk of aGVHD was higher when compared to recipients of single UCB unit (Brunstein et al, 2007). In addition, HLA compatibility also plays a role in aGVHD incidence. Several analyses showed a higher incidence of grade III-IV aGVHD in recipients of 4/6 matched UCBT as compared to 5–6/6 matched UCBT (Rubinstein et al, 1998). Eight patients were evaluable for chronic GVHD. Among these six developed chronic GVHD, three with an extensive disease.

With a median follow-up of 23 months (range 10–27 months), seven patients were alive with a predicted 2 year overall survival (OS) of 50 ± 13%. According to disease category, the predicted 2 year OS for patients with acquired BMFS (SAA and PNH) and inherited BMFS (FA and DC) was 80 ± 17% and 33 ± 16%, respectively. The 2-year OS for patients who were transplanted twice was 50 ± 13%. Seven patients died at 6 months (three from fungal infections, two of aGVHD, one from gram-negative sepsis and one from thrombotic microangiopathy). Among them, five patients experienced primary graft failure.

In conclusion, dUCBT was a feasible and effective strategy to salvage around half of these high-risk patients, taking into account a higher risk of GVHD reported in inherited BMFS. Despite the heterogeneity of the cohort, our results are encouraging in acquired SAA. However, this approach remains of uncertain value in inherited bone marrow failure based on these results. Larger prospective studies are needed to justify the inclusion of dUCBT in the treatment strategy of BMFS.


The work has been supported by an unrestricted grant from EBMT/AMGEN fellowship programme.


Contribution: AR, RPL contributed equally to the study. AR, RPL, VR, EG and GS designed the study, performed research, collected, analysed interpreted data and wrote the paper. MR, CAR, JL, RT, PR, CF, AD took care of patients and reviewed draft of the paper.

Conflict of interest

The authors declare no competing financial interest.