Acute Leukemia

Outcome after failure of allogeneic hematopoietic stem cell transplantation in children with acute leukemia: a study by the société Francophone de greffe de moelle et de thérapie cellulaire (SFGM-TC)

Article metrics


Allogeneic hematopoietic stem cell transplantation (SCT) contributes to improved outcome in childhood acute leukemia (AL). However, therapeutic options are poorly defined in the case of post-transplantation relapse. We aimed to compare treatment strategies in 334 consecutive children with acute leukemia relapse or progression after SCT in a recent 10-year period. Data could be analyzed in 288 patients (157 ALL, 123 AML and 8 biphenotypic AL) with a median age of 8.16 years at transplantation. The median delay from first SCT to relapse or progression was 182 days. The treatment consisted of chemotherapy alone (n=108), chemotherapy followed by second SCT (n=70), supportive/palliative care (n=67), combination of chemotherapy and donor lymphocyte infusion (DLI; n=30), or isolated reinfusion of donor lymphocytes (DLI; n=13). The median OS duration after relapse was 164 days and differed according to therapy: DLI after chemotherapy=385 days, second allograft=391 days, chemotherapy=174 days, DLI alone=140 days, palliative care=43 days. A second SCT or a combination of chemotherapy and DLI yielded similar outcome (hazard ratio (HR)=0.85, P=0.53) unlike chemotherapy alone (HR=1.43 P=0.04), palliative care (HR=4.24, P<0.0001) or isolated DLI (HR=1,94, P<0.04). Despite limitations in this retrospective setting, strategies including immunointervention appear superior to other approaches, mostly in AML.


Hematopoietic stem cell transplantation (SCT) has contributed to improved outcome for pediatric patients with acute leukemia (AL). However, post-transplantation relapse is still associated with a dismal prognosis and its optimal treatment remains unclear. Therapeutic treatments for patients who relapse after first SCT are limited. Treatment options include supportive care, chemotherapy,1, 2 second SCT using the same or an alternative donor3, 4, 5, 6, 7, 8 and more recently donor lymphocyte infusion (DLI).9, 10, 11, 12, 13 So far, no standard approach to this difficult clinical problem has been established. Therapeutic strategies may vary according to the delay between transplant and relapse, as well as centers and child’s specific requirements.

We aimed to compare survival according to different treatment strategies, in case of relapse or progression post allogeneic SCT in children with acute leukemia in a recent 10-year period.

Materials and methods

Data collection

We analyzed all consecutive children (<18 years), who received a first allogeneic SCT for ALL or AML from January 2000 to December 2009 and experienced a relapse or progression thereafter. Clinical data were prospectively collected using ProMISe (Project Manager Internet Server), an internet-based data registry system shared by the 33 centers of the SFGM-TC (French Society of Bone Marrow Transplantation and Cell Therapies) who participated in this study.

The study was designed by the SFGM-TC. The study was approved by the ethics committee at each participating institution and was conducted in accordance with the consent of patients.

The primary end point was overall survival (OS) after diagnosis of relapse or progression post first SCT whatever the treatment post relapse was. Failure was defined as hematological relapse (defined by recurrence of blasts in peripheral blood, or infiltration of bone marrow by more than 5% blasts) or progression of the initial leukemia. Secondary tumors were excluded. Secondary end points included OS after (i) second SCT, (ii) chemotherapy alone, (iii) DLI, (iv) chemotherapy followed by DLI or (v) supportive and palliative care. We also assessed the prognostic risk factors for survival after failure post SCT.

Eligibility criteria

Inclusion was based on the following criteria: age <18 years at first allogeneic transplant, diagnosis of acute leukemia (lymphoid acute leukemia or myeloid acute leukemia), relapse or progression after first SCT, available follow-up data, first allogeneic SCT for ALL or AML from January 2000 to December 2009.

Statistical analysis

Descriptive analyses were created for all study variables using means, SDs, medians, ranges and percentages where appropriate. Quantitative variable were categorized if needed using median, terciles or quartiles. Kaplan–Meier analysis was performed to report survival times and associations between progression and variables were tested with log-rank test.14 Overall survival was defined as the time from the date of relapse after first transplantation or non-remission assessment date until the date of death.

We used a Cox proportional-hazards regression model to examine simultaneously the effects of multiple covariates on survival. We first performed univariate Cox analyses and covariates that achieved a significance level of P0.2 were taken to the multivariate model. Tests for interaction between pairs of variables in the final models were also performed. The effect of each variable in these models was assessed with the use of the Wald test and described by the hazard ratio (HR), with 95% confidence interval. Subjects with missing additional therapy after first relapse were excluded from the Cox models. The final models were developed by introducing all variables of interest in a stepwise backward multivariate analysis. We checked the adequacy of all models using graphical techniques. As the proportion of missing data for therapy variable was 13.8%, secondary analysis was performed using multiple imputation. Missing therapy data were imputed randomly 10 times from scores of similar patients with complete data, and the method of multiple imputation15 was applied to the 10 full data sets so created. The outcome (overall survival) was included among the predictors of these missing data. We addressed the question of the relation between time to relapse and post-relapse treatment choice in two ways; first:we entered time to relapse as a covariate in the Cox models explaining overall survival of the population; second:we systematically tested for second-order interaction between time to relapse and post-relapse treatment. The P-values for interaction tests were not significant and around 0.90. So we did not find any interaction and did not include this interaction term in our models.

Statistical analyses were performed with PASW software version 18.0 (SPSS, Chicago, IL, USA).


Between January 2000 and December 2009, 1307 children received a first SCT for acute leukemia. (ALL n=761, AML n=501, biphenotypic AL n=45). Of whom, 334 children (25.5%) relapsed or progressed thereafter. In 46 of them, the post failure data were not available and 288 patients were thus included in the analysis. The median overall survival of the 46 excluded children was 2.6 months versus 5.3 months for the 288 analyzed patients. We did not observe any significant difference in the median time from transplantation to relapse or progression between included and excluded population (182 days and 180 days, respectively). Analyzed patients had a better median follow-up (160 days and 80 days for excluded children). ALL was the most frequent type of leukemia (n=157) (B-cell phenotype, n=127; T-cell phenotype, n=30), followed by AML (n=123) and biphenotypic AL (n=8; Table 1). At transplantation, 236 (82%) patients were in CR (CR1, n=122, CR2 n=98, CR3 n=16), whereas 52 (18%) had a cytologically detectable disease. Donors were matched related siblings (n=103, 36%), mismatched unrelated (n=70, 24%), matched unrelated (n=51, 18%), unrelated without precision (n=46, 16%), mismatched related (n=11, 4%) or syngeneic (n=5, 2%). Stem cell source was bone marrow (n=184, 65%), umbilical cord blood (n=67, 23%) or peripheral stem cells (n=34, 12%). Median time from diagnosis to first SCT was 245 days. Ninety-one percent (n=263) of the children received a myeloablative regimen and 9% (n=25) received reduced intensity conditioning regimen. TBI was performed in 159 children (55%). Acute GvHD after first SCT occurred in 161 patients (56%; grade I n=62 (38%), grade II to IV n=99 (62%)). The median time from first SCT to relapse/progression was 182 days. The treatments for relapse after first SCT consisted in reinduction chemotherapy alone aiming at obtaining a CR (n=108, 37%), chemotherapy followed by second SCT (n=70, 24%), best supportive care including palliative chemotherapy (n=67, 23%), combination of chemotherapy and DLI (n=30 10%), or isolated DLI (n=13, 6%; Table 2). Analysis of the type of reinduction chemotherapy was hampered by the too large number of different regimens.

Table 1 Characteristics of the 288 patients included in the retrospective study
Table 2 Survival according to treatments after failure of first stem cell transplantation

At the time of analysis in December 2014, 12% of children were alive with a median follow-up of 1315 days (ranges 58–4182). The median OS duration after relapse was 164 days among the 288 patients (Figure 1). ALL was associated with a better outcome as compared with AML (P=0.006; Figure 2).

Figure 1

Overall survival after failure of first stem cell transplantation (all treatments considered).

Figure 2

Overall survival for leukemia subtype. There is a statistically significant difference between acute myeloid leukemia and acute lymphoid leukemia (p<0.006).

Prognostic factors

The following variables impacted the outcome: age at first transplant between 5 and 9 (HR=0.64, P<0.002) and GvHD occurrence after the first transplant (HR=0.60 P<0.002) were associated with a better outcome. Relapse rather than progression was also associated with a better prognosis with a HR of 0.42 (P<0.0001; Table 3). There was no impact on outcome from the following variables: gender, conditioning with TBI, stem cell source (Table 3).

Table 3 Cox model univariate analysis

After univariate analysis, a Cox model multivariate analysis on 288 patients showed the impact of the following factors on survival: CR before relapse (HR=0.60, P<0.004). The time interval from transplantation to relapse was also associated with prognosis: when comparing with relapses occurring within the first 90 days, days 90–182, days 183–364 and 365 days yielded HRs of 0.62, 0.45 and 0.29 (P0.0001), respectively.

The median time from first SCT to relapse/progression was 182 days. Relapse before 182 days was associated with a statistically significant higher rate of failure (P0.003) than late relapse (Table 4).

Table 4 Cox model multivariate analysis including therapy strategy and time from transplant to failure

We then analyzed the impact of the therapeutic strategy on the survival of post-transplant relapsed patients: The patients treated by ‘chemotherapy followed by a second allogenic stem cell transplantation’, and the patients treated by ‘chemotherapy and donor lymphocyte infusion’ had a similar outcome (HR=0.85, P=0.53) with a respective 1-year survival rate of 51 and 53.3%. Treatment with chemotherapy alone (HR=1.43, P=0.04), palliative care (HR=4.24, P<0.0001) or DLI alone (HR=1.94, P<0.04) were associated with a lower survival (Table 4, Figure 3). When omitting the missing values, the statistical analysis yielded comparable results, except for the fact that the difference between the second SCT and chemotherapy alone became nonsignificant (Supplementary Table S1). Additional subgroup analyses show that, in AML patients, a second transplant and chemotherapy+DLI yielded similar results, superior to chemotherapy or palliative care. Conversely, patients with ALL fared better when treated either by second transplant, chemotherapy plus DLI or chemotherapy alone (see Supplementary File). This highlights the higher GVL effect in AML than in ALL.

Figure 3

Overall survival according to therapy after post transplant relapse/progression.

We analyzed whether there was any interaction between late relapse and second hematopoietic SCT/DLI by including treatment strategy as a time-dependent covariate. Although there was an imbalance between the therapeutic groups according to the interval between first SCT and failure, we found no significant relation between the delay of post-transplant relapse and the type of treatment received (P=0.90). Thus, we did not find any statistical relationship between time to relapse and post-relapse treatment that might have influenced the HR obtained in Cox models.


Older strategies for post-relapse therapy mainly relied either on a ‘heavy’ approach consisting of reinduction therapy followed by allogeneic second transplant in remission, or on light care including outpatient leukemia chemotherapy or in several cases only support therapies. New treatment options including DLI emerged in the late nineties (11). This is why we aimed at investigating whether immunotherapeutic approaches can yield a prolonged survival without the cost and treatment burden of a second allogeneic transplant.

In this study, at the time of analysis, 40 patients (12%) were alive with a median follow-up of 1315 days (range 58–4182). AL relapse after first allogeneic SCT still has a very poor prognosis as half of the children did not survive more than 150 days following their relapse or progression. Early relapse was associated with a statistically significantly lower outcome (P0.003) than late relapse. When examining the patient’s related factors influencing outcome, this delay is the most important point to consider. The 1-year survival rates after relapse/progression ranged from 12% for the patients having relapsed in the first 6 months, to 56% in those who relapsed more than 1 year after transplant (Table 5). A previous publication studied the outcome of 25 patients under 18 years with recurrent AML after an initial SCT conditioned with a busulfan and cyclophosphamide preparative regimen.16 The Kaplan–Meier estimates of survival at 100 days, 1 year and 10 years were 88%, 56% and 48%, respectively. In addition, patients who received their second SCT <6 months after the first transplantation were at an increased risk of relapse (P<0.03).

Table 5 Survival according to interval between first transplant and relapse/progression

We have then questioned whether the type of therapeutic approach might impact on the survival. Despite evident limitations due to the retrospective setting of our analysis, best treatment consisted of second transplantation or the combination of chemotherapy plus DLI, both yielded similar long-term results. These results must be taken with caution since the chemotherapy+DLI group is composed of only 30 patients. In a recent report of the Berlin/Frankfurt/Muenster Study Group, 93 children with relapsed ALL non-responsive to chemotherapy according to ALL relapse protocols (03/1990–2006/1999)17 were investigated. They were retrospectively assigned to three therapeutic groups. The median survival after curative (intensive polychemotherapies, SCT), palliative chemotherapy or supportive care were 121, 89 and 42 days, respectively (P<0.001). Time point of relapse and treatment strategy after failure were independent predictors of survival duration.

We are aware that the therapeutic strategy is not independent from the time point of relapse after SCT. The earlier the relapse, the less the doctors and families are prone to aggressive treatments. This is highlighted in our study by the imbalance between the therapeutic groups according to the elapsed time between first SCT and failure (Supplementary Table S2). Median time was shorter in the chemotherapy+DLI arm (211 days) than in the second transplant arm (308 days). The following guidelines could be proposed regarding the best results obtained by allogenic approaches in AML patients. For early (6 months) relapses, taking into account the performance status of the child, the alternatives might lie between palliative/supportive care and a mild approach combining azacytidine plus DLI, as described successfully in adults.18 For patients relapsing in the second semester after transplant, reinduction chemotherapy followed by DLI would be preferred and second transplant only proposed to those patients with persistent positive MRD. After 1 year, chemotherapy followed by a second transplant appears as the best strategy ensuring long-term survival hope.

In ALL patients, very early relapses would warrant innovative approaches with bi-specific antibodies or engineered T cells, or palliative care. For early (7–12 months) and late (>12 months) relapses, the preferred choice would be chemotherapy, either alone or followed by DLI. The benefit of a second transplant in patients with ALL appears questionable, even when the relapse occurs late. The high toxicity of the TBI-based conditioning usually used in the first transplant might contribute to explain the high mortality after second transplant in ALL patients. In all instances, a comprehensive information on treatment-related risks and an intense dialog should prevail, thus allowing the parents and the child to participate actively in the choice between several therapeutic options.

Specific genetic predictors and MRD strongly correlate with the treatment outcome19, 20, 21, 22 and may guide risk-stratified therapeutic decisions. The development of highly sensitive MRD techniques standardized for all patients, genetic profiling and identification of other predicting factors are required for better-individualized treatments. Emerging therapies like bi-specific antibodies23, 24 and CAR T cells25, 26 will profoundly modify the landscape of resistant ALLs, but these targeted therapies are still lacking for childhood AML.


  1. 1

    Frassoni F, Barrett AJ, Granena A, Ernst P, Garthon G, Kolb HJ et al. Relapse after allogeneic bone marrow transplantation for acute leukaemia: a survey by the E.B.M.T. of 117 cases. Br J Haematol 1988; 70: 317–320.

  2. 2

    Mortimer J, Blinder MA, Schulman S, Schulman S, Appelbaum FR, Buckner CD et al. Relapse of acute leukemia after marrow transplantation: natural history and results of subsequent therapy. J Clin Oncol 1989; 7: 50–57.

  3. 3

    Radich JP, Sanders JE, Buckner CD, Martin PJ, Petersen FB, Bensinger W et al. Second allogeneic marrow transplantation for patients with recurrent leukemia after initial transplant with total body irradiation containing regimens. J Clin Oncol 1993; 11: 304–313.

  4. 4

    Barrett AJ, Locatelli F, Treleaven JG, Gratwohl A, Szydło R, Zwaan FE . Second transplants for leukaemic relapse after bone marrow transplantation: high early mortality but favourable effect of chronic GVHD on continued remission. A report by the EBMT Leukaemia Working Party. Br J Haematol 1991; 79: 567–574.

  5. 5

    Michallet M, Tanguy ML, Socie G, Thiébaut A, Belhabri A, Milpied N et al. Second allogeneic haematopoietic stem cell trans- plantation in relapsed acute and chronic leukaemias for patients who underwent a first allogeneic bone marrow transplantation: a survey of the Societe Francaise de Greffe de moelle (SFGM). Br J Haematol 2000; 108: 400–407.

  6. 6

    Eapen M, Giralt SA, Horowitz MM, Klein JP, Wagner JE, Zhang MJ et al. Second transplant for acute and chronic leukemia relapsing after first HLA-identical sibling transplant. Bone Marrow Transplant 2004; 34: 721–727.

  7. 7

    Bosi A, Laszlo D, Labopin M, Reffeirs J, Michallet M, Gluckman E et al. Second allogeneic bone marrow transplantation in acute leukemia: results of a survey by the European Cooperative Group for Blood and Marrow Transplantation. J Clin Oncol 2001; 19: 3675–3684.

  8. 8

    Hosing C, Saliba RM, Shahjahan M, Estey EH, Couriel D, Giralt S et al. Disease burden may identify patients more likely to benefit from second allogeneic hematopoietic stemcell transplantation to treat relapsed acute myelogenous leukemia. Bone Marrow Transplant 2005; 36: 157–162.

  9. 9

    Kolb HJ, Schattenberg A, Goldman JM, Hertenstein B, Jacobsen N, Arcese W et al. Graft-versus-leukemia effect of donor lymphocyte transfusions in marrow grafted patients: European Group for Blood and Marrow Transplantation Working Party Chronic Leukemia. Blood 1995; 86: 2041–2050.

  10. 10

    Porter DL, Roth MS, Lee SJ, McGarigle C, Ferrara JL, Antin JH . Adoptive immunotherapy with donor mononuclear cell infusions to treat relapse of acute leukemia or myelodysplasia after allogeneic bone marrow transplantation. Bone Marrow Transplant 1996; 18: 975–980.

  11. 11

    Collins RH Jr, Shpilberg O, Drobyski WR, Porter DL, Giralt S, Champlin R et al. Donor leukocyte infusions in 140 patients with relapsed malignancy after allogeneic bone marrow transplantation. J Clin Oncol 1997; 15: 433–444.

  12. 12

    Levine JE, Braun T, Penza SL, Beatty P, Cornetta K, Martino R et al. Prospective trial of chemotherapy and donor leukocyte infusions for relapse of advanced myeloid malignancies after allogeneic stem-cell transplantation. J Clin Oncol 2002; 20: 405–412.

  13. 13

    Schmid C, Labopin M, Nagler A, Bornhäuser M, Finke J, Fassas A et al. Donor lymphocyte infusion in the treatment of first hematological relapse after allogeneic stem-cell transplantation in adults with acute myeloid leukemia: a retrospective risk factors analysis and comparison with other strategies by the EBMT Acute Leukemia Working Party. J Clin Oncol 2007; 25: 4938–4945.

  14. 14

    Kaplan EL, Meier P . Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457–481.

  15. 15

    Rubin D . Multiple Imputation for Nonresponse in Surveys. Wiley: New York, NY, USA, 1987.

  16. 16

    Meshinchi S, Leisenring WM, Carpenter PA, Woolfrey AE, Sievers EL, Radich JP et al. Survival after second hematopoietic stem cell transplantation for recurrent pediatric acute myeloid leukemia. Biol Blood Marrow Transplant 2003; 9: 706–713.

  17. 17

    Von Stackelberg A, Völzke E, Kühl JS, Seeger K, Schrauder A, Escherich G et al. ALL-REZ BFM Study Group Outcome of children and adolescents with relapsed acute lymphoblastic leukaemia and non-response to salvage protocol therapy: a retrospective analysis of the ALL-REZ BFM Study Group. Eur J Cancer 2011; 47: 90–97.

  18. 18

    Schroeder T, Czibere A, Platzbecker U, Bug G, Uharek L, Luft T et al. Azacitidine and donor lymphocyte infusions as first salvage therapy for relapse of AML or MDS after allogeneic stem cell transplantation. Leukemia 2013; 27: 1229–1235.

  19. 19

    Bader P, Kreyenberg H, Henze GH, Eckert C, Reising M, Willasch A et al. Prognostic value of minimal residual disease quantification before allogeneic stem-cell transplantation in relapsed childhood acute lymphoblastic leukemia: the ALL-REZ BFM Study Group. J Clin Oncol 2009; 27: 377–384.

  20. 20

    Van der Velden VH, van der Sluijs-Geling A, Gibson BE, te Marvelde JG, Hoogeveen PG, Hop WC et al. Clinical significance of flow cytometric minimal residual disease detection in pediatric acute myeloid leukemia patients treated according to the DCOG AnLL97/MRC AML12 protocol. Leukemia 2010; 24: 1599–1606.

  21. 21

    Reaman GH, Smith FO . Childhood Leukemia. Springer: Berlin, Heidelberg, Germany, 2011.

  22. 22

    Rubnitz JE, Inaba H . Childhood acute myeloid leukaemia. Br J Haematol 2012; 159: 259–276.

  23. 23

    Zugmaier G, Gökbuget N, Klinger M, Bargou RC, Topp MS et al. Long-term survival and T-cell kinetics in relapsed/refractory ALL patients who achieved MRD response after blinatumomab treatment. Blood 2015; 126: 2578–2584.

  24. 24

    Schlegel P, Lang P, Zugmaier G, Feuchtinger T, Handgretinger R . Pediatric post-transplant relapsed/refractory B-precursor acute lymphoblastic leukemia shows durable remission by therapy with the T-cell engaging bispecific antibody blinatumomab. Haematologica 2014; 99: 1212–1219.

  25. 25

    Zhang T, Cao L, Xie J, Shi N, Zhang Z, Luo Z et al. Efficiency of CD19 chimeric antigen receptor-modified T cells for treatment of B cell malignancies in phase I clinical trials: a meta-analysis. Oncotarget 2015; 6: 33961–33971.

  26. 26

    Van der Stegen SJ, Hamieh M, Sadelain M . The pharmacology of second-generation chimeric antigen receptors. Nat Rev Drug Discov 2015; 14: 499–509.

Download references


This work was supported by the Scientific Council of the Société Francophone de Greffe de Moelle et Thérapie Cellulaire. We thank Mrs Rachel Tipton for excellent secretary assistance.

Author information

Correspondence to P-S Rohrlich.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies this paper on Bone Marrow Transplantation website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Roux, C., Tifratene, K., Socié, G. et al. Outcome after failure of allogeneic hematopoietic stem cell transplantation in children with acute leukemia: a study by the société Francophone de greffe de moelle et de thérapie cellulaire (SFGM-TC). Bone Marrow Transplant 52, 678–682 (2017) doi:10.1038/bmt.2016.360

Download citation

Further reading