research paper Summary

Transcription

1 research paper Outcome of aplastic anaemia in children. A study by the severe aplastic anaemia and paediatric disease working parties of the European group blood and bone marrow transplant Summary This study analysed the outcome of 563 Aplastic Anaemia (AA) children aged 0 12 years reported to the Severe Aplastic Anaemia Working Party database of the European Society for Blood and Marrow Transplantation, according to treatment received. Overall survival (OS) after upfront human leucocyte antigen-matched family donor (MFD) haematopoietic stem cell transplantation (HSCT) or immunosuppressive treatment (IST) was 91% vs. 87% (P 018). Event-free survival (EFS) after upfront MFD HSCT or IST was 87% vs. 33% (P 0001). Ninety-one of 167 patients (55%) failed front-line IST and underwent rescue HSCT. The OS of this rescue group was 83% compared with 91% for upfront MFD HSCT patients and 97% for those who did not fail IST up-front (P 0017). Rejection was 2% for MFD HSCT and HSCT post-ist failure (P 073). Acute graft-versus-host disease (GVHD) grade II-IV was 8% in MFD graft vs. 25% for HSCT post- IST failure (P < 00001). Chronic GVHD was 6% in MFD HSCT vs. 20% in HSCT post-ist failure (P < 00001). MFD HSCT is an excellent therapy for children with AA. IST has a high failure rate, but remains a reasonable first-line choice if MFD HSCT is not available because high OS enables access to HSCT, which is a very good rescue option. Keywords: aplastic anaemia, children, haematopoietic stem cell transplantation, immunosuppression. Carlo Dufour, 1 Marta Pillon, 2 Gerard Socie, 3 Alicia Rovo, 4 Elisa Carraro, 2 Andrea Bacigalupo, 5 Rosi Oneto, 5 Jakob Passweg, 4 Antonio Risitano, 6 Andre Tichelli, 4 Regis Peffault de Latour, 3 Hubert Schrezenmeier, 7 Britta Hocshmann, 7 Christina Peters, 8 Austin Kulasekararaj, 9 Anja Van Biezen, 10 Sujith Samarasinghe, 11 Ayad Ahmed Hussein, 12 Mouhab Ayas, 13 Mahmoud Aljurf 13 and Judith Marsh 9 1 Clinical and Experimental Haematology Unit, G Gaslini Childrens Hospital, Genova, 2 Paediatric Haemato-Oncology Clinic, University of Padova, Italy, 3 Department of Haematology, Hopital St Louis, Paris, France, 4 Basel University Hospital, Basel, Switzerland, 5 Second Division of Haematology, San Martino Hospital, Genova, 6 Haematology, Department of Clinical Medicine and Surgery, Federico II University of Naples, Italy, 7 Institute for Clinical Transfusion Medicine and Immunogenetics and Department of Transfusion, Medicine University of Ulm, Germany, 8 Paediatric Haemtopoietic Stem Cell Transplantation, St Anna Kinderspital, Vienna, Austria, 9 Department of Haematological Medicine, King s College Hospital/King s College, London, UK, 10 EBMT Data Office, University Medical Centre, Leiden, The Netherlands, 11 Great Ormond Street Children s Hospital, London, UK, 12 Bone Marrow and Stem Cell Transplantation Program, King Hussein Cancer Centre, Amman, Jordan and 13 King Faisal Specialist Hospital & Research Centre, Riyadh, Saudi Arabia Received 22 October 2014; accepted for publication 3 December 2014 Correspondence: Carlo Dufour, Clinical and Experimental Haematology Unit, G.Gaslini Children s Hospital, Largo G Gaslini 5, Genova 16147, Italy. carlodufour@ospedale-gaslini.ge.it ª 2015 John Wiley & Sons Ltd First published online 14 February 2015 doi: /bjh.13297

2 C. Dufour et al Idiopathic aplastic anaemia (AA) is a rare marrow failure disorder in which pancytopenia occurs as a consequence of the inability of bone marrow to produce adequate amounts of mature blood cells, which, most often, is due to an immune attack on haematopoietic stem cells (Young et al, 2006; Dufour et al, 2009). The estimated incidence of AA in Western Countries is 1 2/million population/year and it is reported to be lower in childhood than in older ages (Young, 1996). This makes it difficult to perform large studies on this rare disease in this particular age group and emphasizes the need for robust data to address appropriate management strategies. The database of the Severe Aplastic Anaemia Working Party (SAAWP) of the European Group Blood and Bone Marrow Transplant (EBMT) contains data on a large number of children diagnosed with AA, thus offering the opportunity to investigate this disease in this population. In a recent study, our group evaluated the outcome of AA in adolescence, which was defined as the period of life stretching from 12 to 18 years (Dufour et al, 2014). The present study evaluated the outcome of AA in childhood, defined as the period of life from birth to age 12 years. To this end, we analysed the records of 563 children aged from 0 to <12 years with the following aims: (i) to compare the outcome after front-line immunosuppressive therapy (IST) with that of front line haematopoietic stem cell transplantation (HSCT) from a matched family donor (MFD). (ii) to evaluate the outcome of patients who, after having failed IST, underwent rescue HSCT and to compare them with front-line MDF HSCT and those who did not fail IST (IST with no subsequent transplant). (iii) to evaluate the cumulative incidence (CI) of post-therapy tumours and (iv) to identify prognostic factors that could affect the outcome of the disease. Methods Patients The records of 563 consecutive children, 313 males and 250 females aged 0 12 years, included in the database of the SA- AWP of the EBMT with diagnosis of AA between 1 January 2000 and 31 December 2009, were analysed. Geographical origin was known in 488 patients and was distributed as follows: 383 patients from Europe, 51 from Africa, 51 from the Middle-East, two from Australia and one from Brazil. Median age at diagnosis was 78 (min 001, max 119) years. Two patients had a paroxysmal nocturnal haemoglobinuria (PNH) clone before HSCT. Median follow-up of survivors was 27 years (min 0 max 111 years). Follow-up was different in the three subgroups: in MFD HSCT first line followup was 26 years (range years), in HSCT post-ist it was 41 years (range years) and it was 17 years (range years) in IST. Constitutional marrow failure syndromes were excluded on the basis of clinical findings and available relevant laboratory tests. Given that patients in the database of the SAAWP of the EBMT are reported according to their first-line therapy, either MFD HSCT or IST, we first analysed the outcome after these two treatments. A proportion of patients who underwent IST upfront failed this treatment and underwent transplantation. We than compared the outcome of these patients (hereafter HSCT post-ist failure) with that of patients who received MFD HSCT and that of subjects who responded to IST and did not require subsequent transplant (hereafter IST alone). Thus in the Results section we refer to the following populations: MFD HSCT (patients who received front line HSCT from a MFD), IST upfront (patients who received IST front-line, including subjects who were transplanted after IST failure), IST alone (patients who initially received and did not fail IST as first line therapy, and therefore did not receive a transplant) and HSCT post- IST failure (patients transplanted after having failed IST). Statistic analysis In order to evaluate the effect of treatments on the outcome, the probability of overall survival (OS) and event-free survival (EFS) were estimated. OS was defined as the time from diagnosis to death from any cause. Event-free survival was calculated from the date of diagnosis to the first event or to the last follow-up. Events were death, lack of response, relapse, occurrence of clinical PNH, secondary malignancies and transplant for patients receiving IST front-line (IST). The same events excluding transplantation were considered in the group receiving IST up-front with no subsequent transplantation (IST alone). Patients lost to follow up were censored at the time of their withdrawal. The log rank test (Peto & Peto, 1972) was used for univariate analysis. Multivariate Cox model analysis was planned for variables with a log rank P-value less than 01 in univariate analysis (Kaplan & Meier, 1958). The selection of parameters for investigation was limited by incomplete data set in certain fields, such as conditioning regimen and type of anti-thymocyte globulin (ATG) used in the IST group. Multivariate analysis took into account comparable parameters for patients who underwent HSCT and/or IST to evaluate the potential impact on outcome. Differences in the distribution of various parameters were compared using chi-square or Fisher exact tests as appropriate. All statistical analyses were carried out by using SAS (SAS-PC, version 9.23; SAS Institute Inc., Cary, NC, USA). Cumulative incidences of malignancies occurring posttreatment, chronic and acute graft-versus-host disease (GVHD) and rejection were estimated in the competing risk model, with relapse or death (not related to late malignancies and GVHD, respectively) as the competing event. Cumulative incidence curves were compared using the Gray test (Kalbfleisch & Prentice, 1980; Gray, 1988). P values <005 were considered to be statistically significant. 566 ª 2015 John Wiley & Sons Ltd

3 Aplastic Anaemia in Childhood Table I. Patients receiving immunosuppressive treatment (IST). Treatment Results Patients (n) Front-line IST 167 Treated pre Treated post No further treatment (IST alone) 76 (66 pre-2007, 10 post-2007) Treatments A total of 167 patients received front-line IST, consisting of ATG plus ciclosporin (CsA). ATG was either horse or rabbit, depending on the time of treatment. Horse ATG was withdrawn from the European market in 2007 and so rabbit ATG was the only available product after this date. Of these 167 children, 91 (55%) failed IST as front-line treatment and underwent subsequent HSCT (HSCT post-ist failure) whereas IST was the only received treatment (IST alone) for 76 patients (Table I). In the HSCT post-ist failure group, the donor was matched unrelated for 63% of patients, matched family in 20% cases and mismatched family or Table II. Characteristics of 563 children diagnosed with aplastic anaemia, included in the study. Transplanted population HSCT post-ist failure Front-line MFD HSCT 487 patients 91 patients 396 patients Donor (%) MFD MUD 63 MMFD or MMUD 17 Conditioning (%) Cy alone 30 74* Cy + Flu 30 11* Cy + others 19 11* Cy + Flu + others 12 2* Flu others 7 2* GVHD Prophylaxis (%) CsA + MTX CsA ATG no/yes 33/67 55/45 Others 21 9 HSCT, haematopoietic stem cell transplantation; IST, Immunosuppressive treatment; MUD, matched unrelated donor; MFD, matched family donor; MMFD, mismatched family donor; MMUD, mismatched unrelated donor; Cy, cyclophosphamide; Flu, fludarabine; GVHD, graft-versus-host disease; CsA, ciclosporin; MTX, methotrexate; ATG, Anti-thymocyte globulin. *Due to incomplete data set percentages are calculated from 255/396 patients for whom data were available. Due to incomplete data set percentages are derived from 219/396 patients, for whom data were available (in the MFD HSCT group) and from 80/91 patients for whom data were available (in the HSCT post-ist failure). unrelated in 17% of children. Conditioning regimen was most frequently cyclophosphamide (Cy) alone followed by Cy plus fludarabine (Flu). Irradiation was not used in patients whose data were available. GVHD prophylaxis was with methotrexate (MTX) plus CsA or CsA alone in the majority of patients. ATG was given to 67% of patients (Table II). A total of 396 children received MFD HSCT as front-line treatment. Unfortunately the data set on conditioning regimen and GVHD prophylaxis was incomplete. Based on available data, Cy alone was the prevalent conditioning regimen followed by Cy plus Flu. Irradiation was not present in the available data. GVHD prophylaxis was conducted in the majority of patients with MTX plus CsA or with CsA alone. ATG was given to 45% of patients (Table II). OS and EFS after front line treatment either IST or MFD HSCT The 3-year probability of OS and EFS for the whole population was 90% and 86%, respectively. The 3-year OS was 91% (standard error [SE] 2%) after MFD HSCT vs. 87% (SE 3%) after first-line IST (P = 018; Fig 1A). Of the 167 patients initially treated with IST, 91 (55%) failed this treatment but were successfully rescued after HSCT (OS 83%, Fig 2A). The 3-year EFS was significantly better in the group of patients who underwent first-line MFD HSCT (87%: 2% SE) than in those treated with IST upfront (33%: SE 4%) (P < 00001; Fig 1B), highlighting the high rate of failure of IST as first-line treatment. OS and EFS after HSCT post-ist failure In order to examine the outcome of subjects transplanted after having failed IST upfront, we extracted this group from that of subjects initially treated with IST. Moreover, for a wider evaluation these patients were compared with those who received IST alone and HSCT from MFD as front-line treatment (see Patients section for definitions). The 3-year probability of OS after HSCT post-ist failure was 83%, after MFD HSCT it was 91% and after IST alone it was 97% (P = 0017). Sub-group analysis showed no significant difference between IST alone and MFD HSCT (P = 021) but significantly higher OS of both MFD (P = 002) and IST alone (P = 0047) over HSCT post-ist failure, thus suggesting that the difference seen in overall comparison is due to the better outcome of both MFD and IST alone compared to HSCT after failed IST (Fig 2A). Causes of deaths are reported in Table SI. The 3-year probability of EFS was 81% after HSCT post- IST failure, 87% after MFD HSCT and 85% after IST alone (P = 053; Fig 2B). Patients who failed IST were transplanted from different donor types (see Table II) and with different conditioning ª 2015 John Wiley & Sons Ltd 567

4 C. Dufour et al (A) (B) Fig 1. (A) Probability of 3-year OS (Kaplan Meier method) for the whole population of 563 patients stratified by treatment. IST: patients receiving IST front-line, including subjects who were transplanted after IST failure. (B) Probability of 3-year EFS (Kaplan Meier method) for the whole population of 563 patients stratified by treatment. IST: patients receiving IST front-line, including subjects who were transplanted after IST failure. OS, overall survival; EFS, event-free survival; HSCT, haematopoietic stem cell transplantation; IST, immunosuppressive treatment; MFD, matched family donor. regimens. This, along with small and unbalanced grouping, precludes reliable statistical analysis and outcome comparisons among different subgroups. Effect of source of cells on OS and EFS in transplanted patients In children transplanted up-front from a MFD, using bone marrow (BM) as the source was associated with a 3-year probability of survival of 93% vs. 80% for peripheral blood (PB) (P =<00013). EFS was 89% with BM and 77% with PB (P =<007; Fig S1A and B). There was a trend for higher rates of acute (194%) and chronic (615%) GVHD with PB vs. BM cells (15% for acute and 352 for chronic GVHD) but this was not significant. The results were similar when data for the whole transplanted population (MFD plus HSCT post-ist failure) was analysed (not shown). Effect of severity of disease and of type of ATG in ISTtreated patients In the selected group of patients treated with IST alone, the rarity of events amongst subgroups precluded the analysis of OS and EFS according to disease severity [non-severe AA vs. severe AA (SAA) vs. very severe AA (VSAA) according to International Criteria (F uhrer et al, 2005)]. Unfortunately, our data set was incomplete regarding the type of ATG (horse or rabbit) used. However, within the limited available data, analysis of the outcome after horse or rabbit ATG in the sub-group of patients receiving IST alone (those who did not undergo subsequent transplant) showed that horse ATG resulted in a better EFS over rabbit ATG (Fig S2A). Interestingly, patients treated before 2007 had a significantly better EFS (92%) than those treated after 2007 (43%; P = 00003), when only rabbit ATG was available in Europe (Fig S2B). 568 ª 2015 John Wiley & Sons Ltd

5 Aplastic Anaemia in Childhood (A) (B) Fig 2. (A) Probability of 3-year OS (Kaplan Meier method) for the whole population of 537 patients stratified by treatment. Front-line IST: patients receiving IST upfront, excluding those who were subsequently transplanted. Subgroup analysis: First-line MFD HSCT vs. Front-line IST: P = 021. First-line MFD HSCT vs. HSCT post- failed IST: P = 002. Front-line IST vs. HSCT post-ist failure: P = (B) Probability of 3-year EFS (Kaplan Meier method) for the whole population of 537 patients stratified by treatment. Front-line IST: patients receiving IST upfront, excluding those who were subsequently transplanted. OS, overall survival; EFS, event-free survival; HSCT, haematopoietic stem cell transplantation; IST, immunosuppressive treatment; MFD, matched family donor. Rejection and GVHD The 2-year CI of rejection was 2% in both MFD (SE 08%) and HSCT post-ist failure (SE 2%) (P = 073) No cases of secondary rejection were reported. Acute and chronic GVHD affect the quality of life of survivors after transplant. Therefore we assessed this parameter in patients receiving first line MFD HSCT and HSCT post-failed IST. The day 100 CI of acute GVHD II-IV was 8% (SE 1%) in MFD HSCT and 25% (SE 5%) in HSCT post-ist failure (P < 00001; Fig 3A). Frequency of acute GVHD III-IV was 3% in MFD and 11% in HSCT post-ist failure. The 5-year CI of chronic GVHD was 6% (SE 2%) in MFD and 20% (SE 4%) in HSCT post-ist failure (P < 00001; Fig 3B). Frequency of chronic GVHD was 4% in MFD HSCT (extensive in 2%) and of 18% in HSCT post-failed IST (extensive in 9%). Post-therapy malignancies After a median follow-up of 27 years, a total of four posttherapy malignancies [two leukaemias, one lymphoma and one myelodysplastic syndrome (MDS)] occurred in 563 children with an overall frequency of 07%. Three malignancies occurred after HSCT post-ist failure, one after MFD HSCT and none after IST alone. The interval between transplant and clonal event was 04 years for the lymphoma, 14 years for the MDS and 28 and 91 years for the two leukaemias. The 5-year CI of malignancies was 03% (SE 03%) in MFD HSCT and 3% (SE 2%) in HSCT post-ist failure (P = 0057). Neoplasms occurred in patients who experienced rejection, indicating the likely origin in recipient cells. ª 2015 John Wiley & Sons Ltd 569

6 C. Dufour et al (A) (B) Fig 3. (A) Cumulative Incidence of acute GVHD grade II-IV in patients transplanted front-line from MFD and post-ist failure. (B) Cumulative Incidence of chronic GVHD in patients transplanted front-line from MFD and after failed IST. Pr, probability; SE, standard error; agvhd, acute graft-versus-host disease; cgvhd, graft-versus-host disease; HSCT, haematopoietic stem cell transplantation; IST, immunosuppressive treatment; MFD, matched family donor. Prognostic factors In the entire population of 563 children, none of the assessed variables (severity of disease, interval diagnosis-therapy 2 months, type of treatment) were shown to affect the outcome in either univariate or multivariate analysis. Within the population of transplanted patients (MFD HSCT plus HSCT post-ist failure) the only factor that was significantly associated, in both univariate and multivariate analysis, with a worse OS and EFS was the use of PB as a source of stem cells (Table IIIA). This was also confirmed only for OS in the population of patients who received MDF HSCT as first-line treatment (Table IIIB). Discussion To our knowledge this is the largest study investigating the outcomes of AA in children aged between 0 and 12 years. Our analysis shows an excellent OS and EFS rate of 90% and 86% respectively that are superior to those we found in adolescents (85% and 78%) (Dufour et al, 2014). 570 ª 2015 John Wiley & Sons Ltd

7 Aplastic Anaemia in Childhood Table III. OS and EFS (A) in all patients who received an HSCT upfront or post-ist failure (n = 448). (B) OS and EFS in patients receiving MFD HSCT upfront (n = 374). Univariate Cox regression Parameter P P Relative risk 95% Confidence limits A OS PB EFS PB B OS PB EFS PB OS, overall survival; EFS, event-free survival; HSCT, haematopoietic stem cell transplantation; IST, immunosuppressive treatment; MFD, matched family donor; PB, peripheral blood. In our study MFD HSCT provides a similar OS but a far better EFS vs. front-line IST. This finding was also observed by a Japanese group (Yoshida et al, 2014). In addition, HSCT is known to provide more durable and powerful haematopoietic reconstitution (Podesta et al, 1998), which is important to reduce the risk of partial haematological recovery often seen after IST, that may lead to unsatisfactory quality of life in children. Moreover GVHD, a complication affecting the quality of life of transplant survivors, occurred only in 3 4% of patients. All these findings indicate that HSCT should be chosen as first-line treatment, if a MFD is available. Regarding IST, even with the paucity of available data, rabbit ATG resulted in an inferior EFS vs. horse ATG. In keeping with this finding, EFS was inferior in patients treated after 2007, when only rabbit ATG was available in Europe vs. prior to 2007 when horse ATG was the preferred source. These findings are in line with two recent prospective studies (Scheinberg et al, 2011; Marsh et al, 2012) that showed a clear superiority of horse over rabbit ATG in upfront IST for AA, hence the current recommendations (Marsh et al, 2009; Samarasinghe & Webb, 2012; Dufour et al, 2013; Samarasinghe et al, 2014) to use horse ATG in front-line IST. Event-free survival after front-line IST was largely unsatisfactory but, similar to another recent study (Jeong et al, 2014), OS was very good (87%). This is very important because survival enables back-up treatment options. In this respect HSCT after failed IST provided an OS of 83% and an EFS of 81%, which are similar to first-line MFD transplants. Even if there was a higher risk of acute and chronic GVHD, possibly reflecting the high proportion of matched unrelated donor (MUD) and mismatched donor grafts, HSCT post-ist failure looks a very good rescue option that still renders front-line IST a reasonable choice if a MFD is not available. This and other (Kosaka et al, 2008) favourable outcomes of HSCT post-ist failure and the knowledge regarding incomplete haematopoietic reconstitution after IST (Podesta et al, 1998) supports current recommendations (Samarasinghe & Webb, 2012; Korthof et al, 2013) that a MUD HSCT should be performed in patients who fail IST, rather than a second course of IST, the outcome of which (30 70% of success) (Di Bona et al, 1999; Scheinberg et al, 2006; Marsh et al, 2009; Scheinberg & Young, 2012) is inferior to that of HSCT post-ist failure shown in the present study. This implies that the search of a donor should be started soon after diagnosis in patients lacking a MFD. The role of upfront MUD HSCT for SAA in children who lack a MFD is much debated issue. Results of MUD HSCT have improved dramatically, and when used after failure of one course of IST, outcomes are similar to MFD HSCT (Kennedy-Nasser et al, 2006). In children receiving reduced intensity conditioning with Flu, Cy and alemtuzumab (FCC), OS and failure free survival exceeded 90% (Samarasinghe et al, 2012). Similar results have been reported in adults with the same conditioning regimen (Eapen et al, 2011). A recent UK national multicentre study of 28 children with SAA/ VSAA who lacked a MFD and were treated with upfront MUD HSCT using FCC, provided excellent OS and EFS (Bhatnagar et al, 2014). This led both the UK Children s Cancer Leukaemia group (CCLG) and EBMT SAAWP to recommend that if a MUD can be found quickly, then HSCT may be considered as upfront treatment in children who lack a MFD (A. Sureda, S. Cesaro, P. Deger, R. Duarte, F. Fakenburg, D. Farge-Bancel, A. Gennery, N. Kroger, F. Lanza, J.C. Marsh, M. Mohty, C. Peters, A. Velardi and A. Madrigal. unpublished observations). However results of controlled studies comparing front-line MUD with other options are awaited to evaluate this new proposed algorithm for treatment of SAA. Similar to other studies (Schrezenmeier et al, 2007; Eapen et al, 2011; Bacigalupo et al, 2012; Dufour et al, 2014) also in childhood, PB resulted in a poorer OS and EFS compared with BM when using ATG-based conditioning and proved to be associated with worse OS and EFS in Cox regression analysis. Therefore BM should be used as the preferred source of cells for transplantation in children with AA. Our study has limits and strengths. The limits include the retrospective nature and the incomplete dataset regarding descriptive items (conditioning regimens and GVHD prophylaxis in the MFD HSCT group). However, the latter does not ª 2015 John Wiley & Sons Ltd 571

8 C. Dufour et al affect the main end-points of the study, which are the outcome of AA in children according to different first and second line treatments, the evaluation of late tumours and the identification of prognostic factors. Another limitation is that a Transplant Registry like ours may generate a bias, preferentially favouring the inclusion of patients that failed IST and subsequently treated with SCT compared to patients that are successfully treated with IST. This may render the study populations somewhat selected. The main strengths are the large number of patients, accounting for the largest cohort of children with AA ever studied, and the streamlined study design enabling robust conclusions to be drawn. In conclusion, this study shows that AA in childhood has an excellent outcome. It also provides evidence to enable recommendations on treatment options that take into account the quantitative and qualitative aspects of survival. If a MFD is available, HSCT using BM cells is the first choice treatment. If a MFD is not available, IST with ATG plus CsA still looks a reasonable front-line option; this is because, despite a relevant rate of failure, the OS is very high and allows access to the very good back-up alternative represented by HSCT post-ist failure. The option of upfront MUD HSCT, if a MFD is not available and a donor is rapidly found, can be considered but needs to be further validated. Authorship and disclosures C Dufour designed the study, elaborated and analysed the data and wrote the paper, M Pillon and E Carraro performed the statistical analysis, elaborated the data, drew the figures and reviewed the paper. R Oneto and A Van Biezen collected and elaborated the data and reviewed the paper, J Marsh, A Bacigalupo, G Socie, J Passweg, H Schrezenmeier, B Hochsmann, M Aljurf, A Kulasekararaj, S Samarasinghe, A Tichelli, AA Hussein, AM Risitano, M Ayas, A Rovo, R Peffault de Latour, C Peters contributed to the study design and data elaboration and reviewed the paper. CD provided consultation for Pfizer. The authors have no competing interests. Acknowledgements ERG spa, Rimorchiatori Riuniti, Cambiaso & Risso, SAAR Depositi Oleari Portuali, UC Sampdoria are acknowledged for supporting the Clinical and Experimental Haematology Unit of G. Gaslini Institute. No specific funding was received for this study. Presented in part at the American Society of Hematology annual meeting, December 2012, Atlanta, GE. Supporting Information Additional Supporting Information may be found in the online version of this article: Fig S1 (A) Probability of 3-years OS (Kaplan Meier method) in patients who underwent upfront HSCT from MFD stratified by source of cells. (B) Probability of 3-years EFS (Kaplan Meier method) in patients who underwent upfront HSCT from MFD stratified by source of cells. Fig S2 (A) Probability of 3-years EFS (Kaplan Meier method) in patients who receive IST alone stratified by ATG type. (B) Probability of 3-years EFS (Kaplan Meier method) in patients who receive IST alone stratified by IST therapy date beginning. Table S1 Causes of death. References Bacigalupo, A., Socie, G., Schrezenmeier, H., Tichelli, A., Locasciulli, A., Fuehrer, M., Risitano, A.M., Dufour, C., Passweg, J.R., Oneto, R., Aljurf, M., Flynn, C., Mialou, V., Hamladji, R.M. & Marsh, J.C. (2012) Bone marrow versus peripheral blood sibling transplants in acquired aplastic anemia: survival advantage for marrow in all age groups. Haematologica, 97, Bhatnagar, N., Wynn, R., Velangi, M., Vora, A., Bonney, D., Gibson, B., Skinner, R., Ewins, A.M., Amrolia, P., Hough, R., De La Fuente, J., Shaw, A., Steward, C., Veys, P. & Samarasinghe, S. (2014) Upfront matched and mismatched unrelated donor transplantation in paediatric idiopathic severe aplastic anaemia: a United Kingdom multicentre retrospective experience. Bone Marrow Transplantation, 49, S2-S3. Di Bona, E., Rodeghiero, F., Bruno, B., Gabbas, A., Foa, P., Locasciulli, A., Rosanelli, C., Camba, L., Saracco, P., Lippi, A., Iori, A.P., Porta, F., De Rossi, G., Comotti, B., Iacopino, P., Dufour, C. & Bacigalupo, A. (1999) Rabbit antithymocyte globulin (r-atg) plus cyclosporine and granulocyte colony stimulating factor is an effective treatment for aplastic anaemia patients unresponsive to a first course of intensive immunosuppressive therapy. Gruppo Italiano Trapianto di Midollo Osseo (GITMO). British Journal of Haematology, 107, Dufour, C., Ferretti, E., Bagnasco, F., Burlando, O., Lanciotti, M., Ramenghi, U., Saracco, P., Van Lint, M.T., Longoni, D., Torelli, G.F., Pillon, M., Locasciulli, A., Misuraca, A., La Spina, M., Bacigalupo, A., Pistoia, V., Corcione, A. & Svahn, J. (2009) Changes in cytokine profile pre- and post-immunosuppression in acquired aplastic anemia. Haematologica, 94, Dufour, C., Svahn, J. & Bacigalupo, A. (2013) Front-line immunosuppressive treatment of acquired aplastic anemia. Bone Marrow Transplantation, 48, Dufour, C., Pillon, M., Passweg, J., Socie, G., Bacigalupo, A., Franceschetto, G., Carraro, E., Oneto, R., Risitano, A.M., Peffault de Latour, R., Tichelli, A., Rovo, A., Peters, C., Hoechsmann, B., Samarasinghe, S., Kulasekararaj, A.G., Schrezenmeier, H., Aljurf, M. & Marsh, J. (2014) Outcome of aplastic anemia in adolescence. A survey of the Severe Aplastic Anemia Working Party of the European Group for Blood and Marrow Transplantation. Haematologica, 99, Eapen, M., Le Rademacher, J., Antin, J.H., Champlin, R.E., Carreras, J., Fay, J., Passweg, J.R., Tolar, J., Horowitz, M.M., Marsh, J.C. & Deeg, H.J. (2011) Effect of stem cell source on outcomes after unrelated donor transplantation in severe aplastic anemia. Blood, 118, F uhrer, M., Rampf, U., Baumann, I., Faldum, A., Niemeyer, C., Janka-Schaub, G., Friedrich, W., Ebell, W., Borkhardt, A. & Bender-Goetze, C. (2005) Immunosuppressive therapy for aplastic anemia in children: a more severe disease predicts better survival. Blood, 106, Gray, R.J. (1988) A class of K-sample tests for comparing the cumulative incidence of a competing risk. Annals of Statistics, 16, Jeong, D.C., Chung, N.G., Cho, B., Zou, Y., Ruan, M., Takahashi, Y., Muramatsu, H., Ohara, A., 572 ª 2015 John Wiley & Sons Ltd

9 Aplastic Anaemia in Childhood Kosaka, Y., Yang, W., Kim, H.K., Zhu, X. & Kojima, S. (2014) Long-term outcome after immunosuppressive therapy with horse or rabbit anti-thymocyte globulin and cyclosporine for severe aplastic anemia in children. Haematologica, 99, Kalbfleisch, J.D. & Prentice, R.L. (1980) The Statistical Analysis of Failure Time Data, pp Wiley, New York, NY. Kaplan, E.L. & Meier, P. (1958) Nonparametric estimation from incomplete observation. Journal of American Statistical Association, 53, Kennedy-Nasser, A.A., Leung, K.S., Mahajan, A., Weiss, H.L., Arce, J.A., Gottschalk, S., Carrum, G., Khan, S.P., Heslop, H.E., Brenner, M.K., Bollard, C.M. & Krance, R.A. (2006) Comparable outcomes of matched related and alternative donor stem cell transplantation for pediatric severe aplastic anemia. Biology of Blood and Marrow Transplantation, 12, Korthof, E.T., BeKassy, A.N. & Hussein, A.A. (2013) Management of acquired aplastic anemia in children. Bone Marrow Transplantation, 48, Kosaka, Y., Yagasaki, H., Sano, K., Kobayashi, R., Ayukawa, H., Kaneko, T., Yabe, H., Tsuchida, M., Mugishima, H., Ohara, A., Morimoto, A., Otsuka, Y., Ohga, S., Bessho, F., Nakahata, T., Tsukimoto, I. & Kojima, S. (2008) Prospective Multicenter trial comparing repeated immunosuppressive therapy with stem-cell transplantation from an alternative donor as second-line treatment for children with severe and very severe aplastic anemia. Blood, 111, Marsh, J.C., Ball, S.E., Cavenagh, J., Darbyshire, P., Dokal, I., Gordon-Smith, E.C., Keidan, J., Laurie, A., Martin, A., Mercieca, J., Killick, S.B., Stewart, R. & Yin, J.A. (2009) Guidelines for the diagnosis and management of aplastic anaemia. British Journal of Haematology, 147, Marsh, J.C., Gupta, V., Lim, Z., Ho, A.Y., Ireland, R., Hayden, J., Potter, V., Koh, M.B., Islam, M.S., Russell, N., Marks, D.I., Mufti, G.J. & Pagliuca, A. (2011) Alemtuzumab with fludarabine and cyclophosphamide reduces chronic graft versus host disease after allogeneic stem cell transplantation for acquired aplastic anemia. Blood, 118, Marsh, J.C., Bacigalupo, A., Schrezenmeier, H., Tichelli, A., Risitano, A.M., Passweg, J.R., Killick, S.B., Warren, A.J., Foukaneli, T., Aljurf, M., Al-Zahrani, H.A., H ochsmann, B., Schafhausen, P., Roth, A., Franzke, A., Brummendorf, T.H., Dufour, C., Oneto, R., Sedgwick, P., Barrois, A., Kordasti, S., Elebute, M.O., Mufti, G.J. & Socie, G. (2012) Prospective study of rabbit antithymocyte globulin and cyclosporine for aplastic anemia from the EBMT Severe Aplastic Anemia Working Party. Blood, 119, Peto, R. & Peto, J. (1972) Asymptotically efficient rank invariant test procedures. Journal of the Royal Stastistical Society Series A, 135, Podesta, M., Piaggio, G., Frassoni, F., Pitto, A., Zikos, P., Sessarego, M., Abate, M., Teresa Van Lint, M., Berisso, G. & Bacigalupo, A. (1998) The assessment of the hematopoietic reservoir after immunosuppressive therapy or bone marrow transplantation in severe aplastic anemia. Blood, 91, Samarasinghe, S., Webb, D.K. (2012) How I manage aplastic anaemia in children. British Journal of Haematology, 157, Samarasinghe, S., Steward, C., Hiwarkar, P., Saif, M.A., Hough, R., Webb, D., Norton, A., Lawson, S., Qureshi, A., Connor, P., Carey, P., Skinner, R., Vora, A., Pelidis, M., Gibson, B., Stewart, G., Keogh, S., Goulden, N., Bonney, D., Stubbs, M., Amrolia, P., Rao, K., Meyer, S., Wynn, R. & Veys, P. (2012) Excellent outcome of matched unrelated donor transplantation in paediatric aplastic anaemia following failure with immunosuppressive therapy: a United Kingdom multicentre retrospective experience. British Journal of Haematology, 157, Samarasinghe, S., Marsh, J. & Dufour, C. (2014) Immune suppression for childhood acquired aplastic anaemia and myelodysplastic syndrome: where next? Haematologica, 99, Scheinberg, P. & Young, N.S. (2012) How I treat acquired aplastic anemia. Blood, 120, Scheinberg, P., Nunez, O. & Young, N.S. (2006) Retreatmet with rabbit anti-thymocyte globulin and cyclosporin for patients with relapsed or refractory severe aplastic anaemia. British Journal of Haematology, 133, Scheinberg, P., Nunez, O., Weinstein, B., Scheinberg, P., Biancotto, A., Wu, C.O. & Young, N.S. (2011) Horse versus rabbit antithymocyte globulin in acquired aplastic anemia. New England Journal of Medicine, 365, Schrezenmeier, H., Passweg, J.R., Marsh, J.C., Bacigalupo, A., Bredeson, C.N., Bullorsky, E., Camitta, B.M., Champlin, R.E., Gale, R.P., Fuhrer, M., Klein, J.P., Locasciulli, A., Oneto, R., Schattenberg, A.V., Socie, G. & Eapen, M. (2007) Worse outcome and more chronic GVHD with peripheral blood progenitor cells than bone marrow in HLA-matched sibling donor transplants for young patients with severe acquired aplastic anemia. Blood, 110, Yoshida, N., Kobayashi, R., Yabe, H., Kosaka, Y., Yagasaki, H., Watanabe, K.I., Kudo, K., Morimoto, A., Ohga, S., Muramatsu, H., Takahashi, Y., Kato, K., Suzuki, R., Ohara, A. & Kojima, S. (2014) First-line treatment for severe aplastic anemia in children: bone marrow transplantation from a matched family donor vs. immunosuppressive therapy. Haematologica, 96, Young, N.S. (1996) Epidemiology of aplastic anemia. In: Bone Marrow Failure Syndromes, 1st edn. (ed. by N.S. Young), pp WB Saunders Company, Philadelphia. Young, N.S., Calado, R.T. & Scheinberg, P. (2006) Current concepts in the pathophysiology and treatment of aplastic anemia. Blood, 108, ª 2015 John Wiley & Sons Ltd 573