Unrelated Donor Searches Impact of marrow unrelated donor ... - Nature

Bone Marrow Transplantation (2003) 32, 325–331 & 2003 Nature Publishing Group All rights reserved 0268-3369/03 $25.00

www.nature.com/bmt

Unrelated Donor Searches Impact of marrow unrelated donor search duration on outcome of children with acute lymphoblastic leukemia in second remission G Dini1, M Grazia Valsecchi2, C Micalizzi1, A Busca3,16, A Balduzzi4, W Arcese5, S Cesaro6, A Prete7, M Rabusin8, E Mazzolari9, P Di Bartolomeo10, N Sacchi11, A Pession7, G Giorgiani12, E Lanino1, T Lamparelli13, C Favre14, A Bosi15, C Manzitti1, S Galimberti2 and F Locatelli12 1 UO Ematologia ed Oncologia Pediatrica, Istituto G Gaslini, Genova, Italy; 2Sezione di Statistica Medica, Dipartimento di Medicina Clinica, Prevenzione e Biotecnologie Sanitarie, Universita` di Milano-Bicocca, Italy; 3Dipartimento Scienze Pediatriche e Adolescentologia, Ospedale Regina Margherita Universita` di Torino, Italy; 4Clinica Pediatrica, Universita` di Milano Bicocca, Ospedale S Gerardo Monza, Italy; 5Dipartimento Biotecnologie Cellulari ed Ematologia, Universita` La Sapienza Roma, Italy; 6 Clinica Oncoematologica Pediatrica, Dipartimento di Pediatria, Universita` di Padova, Italy; 7Dipartimento Scienze Pediatriche Mediche e Chirurgiche, Ospedale S Orsola Malpighi, Bologna, Italy; 8Clinica Pediatrica, Universita` di Trieste, Italy; 9Clinica Pediatrica, Universita` di Brescia, Italy; 10Unita` di Terapia Intensiva Ematologica per il Trapianto Emopoietico, Ospedale Civile di Pescara, Italy; 11Italian Bone Marrow Donor Registry, Italy; 12Oncoematologia Pediatrica, IRCCS Policlinico S Matteo, Universita` di Pavia, Italy; 13Divisione Ematologia II, Ospedale Civile Genova, Italy; 14Clinica Pediatrica, Ospedale S Chiara Universita` di Pisa, Italy; and 15Cattedra di Ematologia, Universita` di Firenze, Italy

Summary: We analyzed the outcome of 167 consecutive children with second CR acute lymphoblastic leukemia (ALL), for whom an unrelated donor (UD) search was activated between 1989 and 1998 at a median time of 2 months after relapse. A suitable donor was identified for 70 patients at 1 year and 6.5 months before and after 1995 from search activation, respectively; a further leukemia relapse occurred during the search in 94 children at a median of 4 months after search activation, 36 of whom underwent UD (14) or other types of transplant (22), beyond second CR, while 58 died of progressive disease. Of 73 patients not experiencing a second relapse, 64 underwent UD (46) or other types of transplant (18), while nine proceeded with chemotherapy, and only four of them survived. The 3-year disease-free survival (DFS) from second CR for the 167 patients is 15.1%, whereas 3-year DFS after transplant for the 60 UD and 40 alternative donor transplanted children is 31.6 and 25.4%, respectively. In conclusion, a further relapse is the main factor adversely affecting outcome of children with second CR ALL. Thus, for these patients, the search should be activated early after relapse and either a UD or an alternative transplant should be performed as early as possible. Bone Marrow Transplantation (2003) 32, 325–331. doi:10.1038/sj.bmt.1704132

Correspondence: Dr G Dini, Department of Pediatric Haematology and Oncology, G Gaslini Children’s Research Hospital, Largo G Gaslini 5, 16148 Genova, Italy 16 Current address: UOA Ematologia, Ospedale San Giovanni Battista, Torino, Italy Received 23 December 2002; accepted 20 February 2003

Keywords: childhood acute lymphoblastic leukemia; unrelated marrow donor; bone marrow transplantation; cord blood transplantation

Despite the progress achieved over the last 20 years in the treatment of childhood acute lymphoblastic leukemia (ALL), patients who experience a relapse still have dismal prognoses when treated with conventional chemotherapy.1 Stem cell transplantation (SCT) from a matched related donor cures more than 50% of patients who failed first-line chemotherapy,1,2 whereas indications for autologous transplantation are limited to a small subset of patients with either late marrow relapse or extramedullary recurrence.3–5 Over the past decade, bone marrow transplantation (BMT) from an unrelated donor (UD) has become a feasible procedure able to cure a significant number of children with ALL lacking an HLA identical family donor.6–11 HLA polymorphism is still a major obstacle in finding a fully matched UD, despite the availability of more than 7 million volunteer UDs in worldwide registries.12 As a result of this, several transplant centers have recently offered transplantation from alternative donors, either HLAmismatched relatives13 or unrelated umbilical cord blood (CB) donors.14 So far, to our knowledge, all published studies have reported the outcome of selected cohorts of patients with ALL who actually received a UD BMT.6–11 In this paper, we provide the full picture regarding all patients for whom a search for a UD was activated through the Italian Bone Marrow Donor Registry (IBMDR). Specifically, the aims of this study were: (1) to evaluate the impact of SCT timing on the outcome of 167 unselected Italian children with ALL in second CR searching for a UD and (2) to assess the probability of finding an HLA-matched UD between 1989 and 1998 through the international registries.

URD search in relapsing childhood ALL G Dini et al

326

Patients and methods Patients Between June 1989 and April 1998, 179 UD searches were activated through the IBMDR for children with ALL who experienced a first bone marrow relapse and were followed by Associazione Italiana di Ematologia ed Oncologia Pediatrica (AIEOP) centers. Their individual data were taken from the IBMDR and the AIEOP registry.6,15,16 All searches were discontinued as of July 1999. The timing of search activation varied over the years: before 1995 it was required that the patient achieved second CR; most centers have only recently activated the search as soon as possible following the first relapse. Of the relapsed patients, 12 did not achieve second remission and were excluded from this analysis. Duration of first CR was less than 30 months in 119 out of 167 evaluable patients (early relapse) and higher than 30 months in the remaining 48 patients (late relapse). All patients included in the study had received front-line treatment according to the BFM-like AIEOP protocols 88, 91, and 9517–19 and had achieved second remission after second-line treatment given according to BFM and AIEOP guidelines (Table 1).20–22

Transplant procedures Starting from 1980, AIEOP and Gruppo Italiano per il Trapianto di Midollo Osseo (GITMO) eligibility criteria for SCT recommended allogeneic transplant from HLAcompatible siblings as the first choice for children in second CR.1,2 More recently, a search for a UD was recommended for patients lacking compatible related donors,4 and this option has become increasingly feasible in later years. All donor searches were facilitated by the IBMDR, as

Table 1

previously described,23 and run on worldwide registries12 after obtaining informed consent from patient’s guardians. The search was discontinued when a UD was found, or when other treatments, including any other types of transplant, were adopted or when the patient died or was considered no longer eligible for an allograft. Before December 1991, most transplant centers (TC) based their matching criteria on HLA-A, B, and DR compatibility as defined by serological typing. After January 1992, all patients were DR subtyped by high-resolution molecular techniques, as previously described.6 Donors who were incompatible for a single HLA locus were considered acceptable if there was a disparity at the HLA-A or B loci for one antigen, or if there was disparity at the DRB1 alleles associated with the same DR serological type.6 Preparative regimen and GvHD prophylaxis for patients undergoing transplant. The preparative regimen for children undergoing UD BMT included 1200 cGy total body irradiation (TBI), in six fractions over 3 days, followed by cyclophosphamide (Cy), 60 mg/kg 2 days, combined with either thiotepa (TT), 2.5 mg/kg/day 2 days, or, in very few cases, with other drugs, as previously reported.24 GvHD prophylaxis varied according to TC and with time. All UDBMT patients received cyclosporine A (CsA) and ‘short methotrexate’ (MTX). This combination was associated with either antithymocyte globulin (ATG) or Campath 1G, as previously reported.24 Preparative regimens and GvHD prophylaxis of the other types of SCT varied over time, according to the source of stem cells, TC and cooperative protocols, as previously reported.3,13,14 Statistical analysis. The Kaplan–Meier estimate and the log-rank test were used to calculate and compare the

Characteristics of patients at search activation, overall and by treatment received UD found (n=70)a

Total n=167 Age at 1st relapse (years) Median Range Gender M/F Duration of 1st CR p30 months >30 months

UD BMT 2nd CR UD BMT >2nd CR Chemo only Other SCT 2nd CRa Other SCT 3rd CR Chemo only (n=46) (n=14) (n=9) (n=18) (n=22) (n=58)

7.6 0.9–16.3 110/57

8.8 1.7–16.2 35/11

6.6 3.6–16.3 9/5

10.7 4.0–15.5 4/5

8.7 4.7–15.3 9/9a

6.3 1.2–12.0 16/6

6.7 0.9–14.8 37/21

119 48

29 17

11 3

8 1

7a 11

17 5

47 11

2 2 days–12

1.6 8 days–5.5

2.3 16 days–5.1

1.6 7 days–4

2.1 13 days–13

2.3 1 day–12.3

5.4 2–14

6.8 3.6–14

4.3 1.1–13

10.6 2.8–58

15 31

2 12

3a 15

9 13

Time from relapse to search activation (months) Median 2.1 Range 1 day–13 Time from search activation to SCT (months) Median Range Period of search activation p1994 58 >1995 109 a

UD not found (n=97)a

1 8

One patient who found a UD underwent another type of SCT (haploidentical). UD BMT, unrelated donor bone marrow transplant; CR, complete remission; SCT, stem cell transplant; M, male; F, female.

Bone Marrow Transplantation

28 30


cumulative probability of events. In detail, the probability of finding a UD took into account the time from search activation to the date when the donor was actually found. Patients were censored at discontinuation of the search for whatever reason, that is, relapse, death, medical or parental decision. The probability of finding a UD in second CR was estimated accounting for relapse, death in CR, and other types of SCT as competing events. The probability of experiencing a second relapse was estimated with the same approach based on competing risk.25 Disease-free survival (DFS) and survival probabilities were calculated as the time from second CR to the first event, that is, relapse or death, whichever occurred first, for DFS, or death due to any cause for survival. The Kaplan– Meier estimate accounts for left truncation, when search is activated after achieving second CR. Date of BMT was considered as time origin for evaluation of DFS and survival after BMT. Follow-up was updated as of 30 June 2001 with a minimum potential follow-up after first relapse of 3.4 years. Software packages used for the analyses were SAS and S-Plus.

Results Search process and outcome The number of search activations for patients in second remission increased markedly during the period that we analyzed. A total of 58 searches were activated before 1995, with an average of 10 searches per year. This average increased to 33 in the following years and reached a total of 109 activations over 40 months. Overall, the median time from first relapse to second CR and to search activation was 1 month (range 15 days–5.6 months) and 2 months (range 1 day–12.8 months), respectively. Most patients (150 over 167) achieved second CR within 2 months and only 50 patients had a search started before achieving second CR. The overall Kaplan–Meier probability of finding a UD within 6 months after search activation for the 167 patients was 33.4% (s.e. 4.2). This improved significantly over time from 22.2% (s.e. 5.9) to 39.9% (s.e. 5.6), before and after 1995, with a median time of 1 year and 6.5 months, respectively (P ¼ 0.02). The IBMDR was the main source of UDs (44.3%), and approximately one of every two donors was Italian. An HLA-compatible donor was identified for 70 patients (36%) at a median time of 5.4 months (range 1.6–15.6 months) after search activation, namely for 19 (32.8%) and 51 (46.8%) who started a search before and after 1995, respectively. The probability of finding a UD for patients in second CR was also evaluated in the light of the disease course. In particular, the occurrence of competing events (relapses, death in CR, other types of SCT) was taken into consideration. Among the 167 patients searching for donors, 94 (56%) relapsed before finding a donor, at a median of 4 months (range 10 days–56 months) from search activation. The cumulative incidence of relapse during search (ie before any type of transplant) was 38.1% (s.e. 3.8) and 49.2% (s.e. 3.9) at 6 and 12 months, respectively. In contrast, the probability of finding a UD

before relapse was 19.5% (s.e. 3.1) and 26.9 (s.e. 3.5) at 6 and 12 months. These data clearly show that the course of the disease during the search, most importantly the occurrence of a second relapse, strongly influences treatment and outcome.

327

Patients not experiencing a second relapse during the search A total of 73 patients did not experience a second relapse during the search (Table 2, column 1). Among them, only nine did not undergo any type of SCT: five had died of infection at a median time of 5.8 months from search activation (range 29 days–1 year), while for the other four patients the search had been discontinued by the treating physicians at a median time of 3.5 months (range 2 months–1.6 years). These four patients are still alive and disease-free. Noticeably, first CR had lasted more than 5 years for three of them. A total of 96 patients underwent UD BMT in second CR at a median time of 5.4 months from search activation (range 2 months–1.3 years), and among them 14 are still alive and disease-free. A total of 17 other patients lacking a UD, and one for whom a suitable UD had been found, underwent other types of SCT, at a median time of 4.3 months (range 1.1 months–1.1 years) after search activation: nine patients underwent autologous bone marrow transplant, five were grafted from a related donor (cord blood in two cases, three antigen mismatched parent in three cases), and four underwent cord blood transplant from a UD (UD CBT). Of these, five are still alive and disease-free. The estimated percent DFS at 3 years after UD BMT was 32.6 (s.e. 6.9) and, since all relapses (n ¼ 12) were followed by death within 3 years, the survival figure was the same. The estimated 3-year percent DFS and survival after

Table 2 Outcome by treatment in patients not experiencing a second relapse during the search (n=73) and in patients experiencing a second relapse (n=94)a Treatment

Chemotherapy only Deaths in CR Deaths with PD Alive in CR UD BMT Deaths in CR Relapses post-BMT (deaths) Alive in CR Autologous SCT Deaths in CR Relapses (deaths) Alive in CR Other types of SCT Deaths in CR Relapses (deaths) Alive in CR

2nd relapse during search No (n=73)

Yes (n=94)

9 5 0 4 46 20 12 (12) 14 9 1 6 (6) 2 9 4 2 (2) 3

58 0 58 0 14 8 2 (2) 4 8 0 4 (2) 4 14 7 6 (6) 1

a

For these latter patients, the reported types of SCT were performed in phases subsequent to second CR. n, number; CR, complete remission; PD, progressive disease; UD, unrelated donor; BMT, bone marrow transplantation; SCT, stem cell transplant. Bone Marrow Transplantation


1.0

Patients experiencing a second relapse during the search

0.8

A total of 94 patients relapsed during the search (Table 2, column 2): 58 of them did not undergo any type of SCT, including nine children with a suitable donor. These latter children, who had relapsed at a median time of 3 months after search activation (range 10 days–6 months), died of progressive disease after few months from relapse (median time being 4 months). Other 49 patients, who relapsed after a median time of 3.6 months (range 18 days–4.2 years) after search activation, and who lacked suitable donors, died of progressive disease after a median time of 6 months. Suitable UDs were located for 14 more patients who had relapsed at a median time of 3.9 months after search activation (range 1.1–11.4 months). The UD BMT was performed 6.8 months after search activation (range 3.6 months–1.2 years) in a more advanced phase of the disease and four children are still alive and disease free. Other 22 patients, who had relapsed at a median time of 6.3 months (range 28 days–4.7 years) and who lacked a UD, were given other types of SCT at a median time of 10.6 months (range 2.8 months–4.9 years) after search activation (eight autologous, 13 from a three antigen mismatched relative, and one from UD CB). Among them, five are still alive and disease free. The estimated 3-year percent DFS after UD BMT was 28.6 (s.e. 13.4) and the 3-year survival figure was the same. The estimated 3-year percent DFS and survival after other types of SCT was 22.9 (s.e. 10.9) and 28.6 (s.e. 10.9), respectively.

Outcome After a median follow-up of 5.4 years, regardless of the treatment received, DFS in the unselected series of 167 patients was 15.1% (s.e. 2.7) at 3 years from second CR. In particular, the subset of 119 patients whose first relapse was ‘early’, that is, within 30 months after diagnosis, had an even poorer prognosis, with a 3-year DFS of 6.4% (s.e. 5.0), as compared to the 48 patients with ‘late’ first relapse, whose DFS was 34.7% (s.e. 7.2) (Po0.001). The ultimate outcome of these series of 167 patients, as described by survival, accounts for the impact of BMT performed after a second relapse and gives a 3-year figure of 24.5% (s.e. 3.3) (Figure 1). The estimated 3-year percent DFS after transplant of the 60 patients given a UD BMT was 31.6% (s.e. 6.0) and the 3-year survival figure was the same (Figure 2). TRM in the 60 UD-BMT patients accounted for 28 of the 42 deaths. The nine patients in our case series who were transplanted more recently (after 1 January 1998) had a lower (although not significantly) 100-day mortality rate as compared to the 51 patients transplanted in earlier years (22.2% (s.e. 13.9) vs 38.7% (s.e. 7.0)). The estimated 3-year percent DFS and survival after transplant for the 40 patients undergoing other types of SCT was 25.4% (s.e. 7.0) and 30.8% (s.e. 7.4), respectively (Figure 3). Bone Marrow Transplantation

Probability

other types of SCT was 27.8 (s.e. 10.6) and 33.3 (s.e. 11.1), respectively.

Survival DFS

0.6

0.4 24.5 (3.3) 20.3 (3.2)

0.2

15.1 (2.7) 10.9 (2.4)

0.0 1

0

3 2 Years from 2nd CR

4

5

Figure 1 Estimated survival and disease-free survival of the 167 patients for whom the search for an unrelated donor was activated.

1.0

0.8

Probability

328

0.6

0.4 31.6% (6.0)

29.7% (6.0)

0.2

0.0 1

0

3

2

4

5

Years from BMT Pts at risk DFS

60

22

19

Survival

60

23

20

17

10

6

17

10

6

Figure 2 Estimated survival and disease-free survival of the 60 patients given an unrelated donor bone marrow transplantation.

Discussion This study shows the history of the entire series of patients with ALL in second remission after a bone marrow relapse, for whom a search for a UD was started in a given period (1989–1998) in Italy. Interestingly, this study gives the full picture of how both the course of the disease and the timing


1.0

Probability

0.8

0.6

0.4 30.8% (7.4)

27.0% (7.4)

25.4% (7.0)

0.2

21.8% (6.9)

0.0 1

0

3

2

4

5

Years from BMT Pts at risk DFS

40

12

11

8

5

5

Survival

40

16

12

9

6

5

Figure 3 Estimated survival and disease-free survival of the 40 patients given other types of stem cell transplantation.

of the search process influence treatment selection and clinical outcome. The series of 167 analyzed patients had, regardless of treatment received, a relatively poor outcome: 24.5% (s.e. 3.3) survival at 3 years from search activation. This outcome may reflect a background negative selection as, in most part of the period considered here, a search for a UD was likely to be started only for patients with a less favorable prognosis. Indeed, many of these patients (56%) had a second relapse very soon during the search process and the majority of them were not in a condition to be transplanted in a subsequent phase. Thus the study includes 67 patients who did not undergo any type of SCT, of whom the only surviving ones (four patients) had an unusually long first remission as compared to other patients in this case series. On the other hand, if we consider the outcome of children who underwent UD BMT, this is comparable to most published reports for children undergoing UD BMT before 1998.7–9 In this study, we were able to evaluate the impact of UD BMT both in second remission and after a second relapse had occurred during the search in 94 patients. The possibility of undergoing UD BMT after a second relapse was very limited due to the fact that most of these patients died from disease progression. Only 14 of the 94 children received this type of transplant. The apparently similar results of these UD BMTs as compared to the outcome of the 46 UD BMTs performed in second CR could be related to favorable patient selection after second relapse. These considerations, together with the data showing that median time to second relapse was 4 months, suggest that UD BMT should be undertaken within 4 months from first relapse. This could be achieved if a search was started at relapse, since our data show that 40% of the searches

activated after 1995 were successful within 6 months after activation and this figure has probably increased more recently thanks to a more efficient system. The length of time required for the search and the intensive salvage chemotherapy administered before transplant may have contributed to the high TRM rate observed in transplants performed before 1998. This suggests that early transplant and better control of early TRM could substantially change the picture. Results recently reported by ourselves and by other groups show that the 2-year probability of DFS after UD BMT is similar to the results obtained with SCT from a matched related donor.26,27 This improvement is mainly due to refinements in HLA typing, GvHD prophylaxis, timing of SCT, and supportive care. It has recently been reported that matching HLA class I and class II alleles of the donor and recipient can significantly improve outcome after UD BMT,28 but, in keeping with Souillet et al,29 the aim of the search strategy should be to carry out a transplant within 3 months from search activation. In the absence of a fully compatible donor, a one-antigen mismatched unrelated volunteer donor can be considered.29 In this study, the outcome of the cohort of patients undergoing other types of SCT should be considered with caution, due to the heterogeneity of transplants and to the small number of analyzed patients. About half of these alternative transplants were performed during a more advanced phase, at a median time of 10.6 months after search activation. Six of 17 patients who underwent ABMT survived, disease-free, but it must be noted that this option was offered mostly to children experiencing a late relapse. This is in keeping with BFM data showing that among 19 children with ALL who underwent ABMT after bone marrow relapse occurring more than 30 months after diagnosis, the 6-year EFS was higher than 50%.30 The international debate on options other than UD BMT is still very open. Recently, more than 70 000 validated CB units have been made available for transplant and have provided more than 2000 UD CBTs around the world. This proves that UD CBT is a feasible procedure that can potentially cure a significant number of children with poor prognosis ALL, especially if transplanted in a favorable phase. The 28% 2-year DFS of 40 children with ALL who underwent UD CBT reported by the Eurocord Registry14 was not different from what was reported in children undergoing a UD BMT. This is why it has been suggested that UD CBT should be taken into consideration for all children with poor prognosis ALL who are eligible for a UD transplant.14 The main advantages of unrelated CBT are both the prompter availability of cryopreserved placental blood and the less stringent requirements for HLA compatibility between donor and recipient. The major drawbacks are the high incidence of transplant-related deaths due to the increased risk of graft failure, and delayed hematological reconstitution. While SCT from phenotypically identical or one-locus mismatched family donors approximates the outcome of patients transplanted from genotypically identical siblings,31 the use of three-loci mismatched family donors is still considered an experimental approach, despite prompt availability of the donor. This view was supported by the

329



330

high TRM. Recently, it has been shown that transplantation of highly purified CD34+ stem cells is a realistic therapeutic option for patients who would otherwise have no suitable donor and who are in need of transplantation.13 In conclusion, our data show that (a) relapse is the major limiting factor for UD BMT in children with ALL in second CR after a bone marrow relapse, and (b) the median time needed to identify a UD significantly decreased in the past 5 years. We conclude that, when UD BMT is a valid therapeutic option, as after an early bone marrow relapse, it is important to perform it within 4 months from relapse. To achieve this goal, UD and CB searches should be initiated during the first CR in patients at high risk for relapse. The same criteria should hold for other transplant options, should their role be better defined in other studies. Finally, a methodological lesson can be learnt from this study: tracking the UD search process for all patients who enter it gives an important insight into the mechanisms that guide treatment allocation. Only in this way the selection biases that are known to operate in studies involving transplant options become evident and help in the interpretation of any treatment comparison.

Acknowledgements We would like to thank data managers from all the centers for their help in collecting data and answering our queries; Roberta Cancedda from G Gaslini Institute (Genova); Francesca Predieri from CO FONOP-AIEOP (Bologna) for their assistance with data management; partially supported by Ricerca Finalizzata 2002, by Compagnia San Paolo, by Fondazione CARIGE.

References 1 Uderzo C, Valsecchi MG, Bacigalupo A et al. Treatment of childhood acute lymphoblastic leukemia in second remission with allogeneic bone marrow transplantation and chemotherapy: ten-year experience of the Italian Bone Marrow Transplantation Group and the Italian Pediatric Hematology Oncology Association. J Clin Oncol 1995; 13: 352–358. 2 Uderzo C, Rondelli R, Dini G et al. High-dose vincristine, fractionated total-body irradiation and cyclophosphamide as conditioning regimen in allogeneic and autologous bone marrow transplantation for childhood acute lymphoblastic leukemia in second remission: a 7-year Italian multicentre study. Br J Hematol 1995; 89: 790–797. 3 Messina C, Cesaro S, Rondelli R et al. Autologous bone marrow transplantation for childhood acute lymphoblastic leukemia in Italy. Bone Marrow Transplant 1998; 21: 1015– 1021. 4 Dini G, Cornish JM, Gadner H et al. Bone marrow transplant indications for childhood leukemias: achieving a consensus. Bone Marrow Transplant 1996; 18 (Suppl. 2): 4–7. 5 Balduzzi A, Gaipa G, Bonanomi S et al. Purified autologous grafting in childhood acute lymphoblastic leukemia in second remission: evidence for long term clinical and molecular remission. Leukemia 2001; 15: 50–56. 6 Dini G, Lanino E, Lamparelli T et al. Unrelated donor marrow transplantation: initial experience of the Italian Bone Marrow Transplant Group (GITMO). Bone Marrow Transplant 1996; 17: 55–62. Bone Marrow Transplantation

7 Balduzzi A, Gooley T, Anasetti C et al. Unrelated donor marrow transplantation in children. Blood 1995; 86: 3247–3256. 8 Casper J, Camitta B, Truitt R et al. Unrelated bone marrow donor transplants for children with leukemia or myelodysplasia. Blood 1995; 85: 2354–2363. 9 Davies SM, Wagner JE, Shu XO et al. Unrelated donor bone marrow transplantation for children with acute leukemia. J Clin Oncol 1997; 15: 557–565. 10 Oakhill A, Pamphilon DH, Potter MN et al. Unrelated donor bone marrow transplantation for children with relapsed acute lymphoblastic leukaemia in second complete remission. Br J Haematol 1996; 94: 574–578. 11 Hongeng S, Krance RA, Bowman LC et al. Outcomes of transplantation with matched-sibling and unrelated donor bone marrow in children with leukemia. Lancet 1997; 350: 767–771. 12 Bone Marrow Donors Worldwide, Edition of June 2001. 13 Aversa F, Tabilio A, Velardi A et al. Transplantation of highrisk acute leukemia with T-cell-depleted stem cells from related donor with one fully mismatched HLA haplotype. N Engl J Med 1998; 339: 1186–1193. 14 Locatelli F, Rocha V, Chastang C et al. Factors associated with outcome after cord blood transplantation in children with acute leukemia. Blood 1999; 93: 3662–3671. 15 Pession A, Rago S, De Rosa M et al. Modello di gestione dei dati: Studio nazionale AIEOP per la leucemia acuta linfoblastica del bambino. Haematologica 1987; 72: 249–254. 16 Pession A, Rondelli R, Paolucci P et al. Hematopoietic stem cell transplantation in childhood: report from the Bone Marrow Transplantation (BMT) group of the Associazione Italiana Ematologia Oncologia Pediatrica (AIEOP). Haematologica 2000; 85: 638–646. 17 Conter V, Arico` M, Valsecchi MG et al. Extended intrathecal methotrexate may replace cranial irradiation for prevention of CNS relapse in intermediate risk ALL children treated with BFM-based intensive chemotherapy. J Clin Oncol 1995; 13: 2497–2502. 18 Vecchi V, Arico` M, Basso G et al. Risk-directed therapy for childhood acute lymphoblastic leukemia. Cancer 1993; 72: 2517–2524. 19 Conter V, Arico` M, Valsecchi MG et al. Intensive BFM chemotherapy for childhood ALL: interim analysis of the AIEOP-ALL 91 study. Haematologica 1998; 83: 791–799. 20 Buhrer C, Hartmann R, Fengler R et al. Superior prognosis in combined compared to isolated bone marrow relapse salvage therapy of childhood acute lymphoblastic leukemia. Med Pediatr Oncol 1993; 21: 470–476. 21 Rossi MR, Masera G, Zurlo MG et al. Randomized multicentric Italian study on two treatment regimens for marrow relapse in childhood acute lymphoblastic leukemia. Pediatr Hematol Oncol 1986; 3: 1–9. 22 Giona F, Testi AM, Annino L et al. Treatment of primary and relapsed acute lymphoblastic leukaemia in children and adults: the GIMEMA/AIEOP experience. Br J Haematol 1994; 86: 55–61. 23 Dini G, Lamparelli T, Rondelli R et al. Unrelated donor marrow transplantation for chronic myelogenous leukaemia. Br J Haematol 1998; 102: 544–552. 24 Locatelli F, Zecca M, Messina C et al. Improvement over time in outcome for children with acute lymphoblastic leukemia in second remission given hematopoietic stem cell transplantation from unrelated donors. Leukemia 2002; 16: 2228–2237. 25 Marubini E, Valsecchi MG. Analysing Survival Data from Clinical Trials and Observational Studies. Wiley: New York, 1995. 26 Gustafsson A, Remberger M, Winiarski J, Ringden O. Unrelated bone marrow transplantation in children: outcome


331 and a comparison with sibling donor grafting. Bone Marrow Transplant 2000; 25: 1059–1065. 27 Saarinen-Pihkala UM, Gustafsson G, Ringdeń O et al. No disadvantage in outcome of using matched unrelated donors as compared with matched sibling donors for bone marrow transplantation in children with acute lymphoblastic leukemia in second remission. J Clin Oncol 2001; 19: 3406–3414. 28 Petersdorf EW, Gooley TA, Anasetti C et al. Optimizing outcome after unrelated marrow transplantation by comprehensive matching of HLA class I and II alleles in the donor and recipient. Blood 1998; 92: 3515–3520.

29 Souillet G, Rey S, Bertrand Y et al. Outcome of unrelated bone marrow donor searches in 174 children resulting in 45 patients transplanted in the HLA-matched and –mismatched situation. Bone Marrow Transplant 2000; 26: 31–43. 30 Borgmann A, Schmid H, Hartmann R et al. Autologous bonemarrow transplants compared with chemotherapy for children with acute lymphoblastic leukaemia in second remission: a matched pair analysis. Lancet 1995; 346: 873–876. 31 Beatty PG, Clift RA, Mickelson EM et al. Marrow transplantation from related donors other than HLA-identical siblings. N Engl J Med 1985; 313: 765–771.
