Pediatr Blood Cancer 2015;62:153–157
Allogeneic Hematopoietic Stem Cell Transplantation in Pediatric Myelodysplastic Syndromes: A Multicenter Experience From Argentina Ana L. Basquiera, MD,1,2* Silvia Pizzi, MD,3 Agustı´n Gonza´lez Correas, MD,3 Pablo G. Longo, MD,4 Wanda C. Goldman, MD,5 Marı´a V. Prates, MD,6 Sandra Formisano, MD,5 Gustavo Kusminisky, MD,4 Leonardo Feldman, MD,7 Adriana R. Berretta, MD,1 Juan J. Garcı´a, MD,1,2 and Raquel Staciuk, MD3 Background. Allogeneic hematopoietic stem cell transplantation (AHSCT) represents the only curative treatment for the majority of pediatric patients with Myelodysplastic Syndrome (MDS). We aimed to evaluate overall survival (OS), disease-free survival (DFS), nonrelapse mortality (NRM) and relapse incidence in children who underwent AHSCT for MDS in six institutions from Argentina. Procedure. A retrospective analysis of 54 AHSCT was carried out in 52 patients (mean age: 9 years; range: 2–19; 35 males). Results. MDS subtypes were refractory cytopenia of childhood (RCC) (n: 26, 50%), refractory anemia with excess blasts (RAEB) (n: 9, 18%), RAEB in transformation (RAEB-T) (n: 8, 15%) and juvenile myelomonocytic leukemia (JMML) (n: 9, 17%). At time of transplant, seven (13%) patients transformed to acute myeloid leukemia (AML) and two
patients with RCC to RAEB. Donors were related in 32 cases (59%) and the stem cells source was: bone marrow (63%), peripheral blood (26%), and umbilical cord blood (11%). Five-year DFS and OS were 50% and 55% respectively; and for patients with JMML, 57% and 67% respectively. Cumulative incidence of NRM and relapse were 27% and 21% respectively. In the multivariate analysis, umbilical cord blood (HR 4.07; P ¼ 0.025) and age 9 years at transplantation (HR 3.28; P ¼ 0.017) were associated with lower OS; age and graftversus-host disease (GVHD) had a higher NRM. Conclusions. In our series, more than half of the patients achieved long term OS with AHSCT. Less toxic conditioning regimens or more intensive GVHD prophylaxis could lead to better results in some children. Pediatr Blood Cancer 2015;62:153–157. # 2014 Wiley Periodicals, Inc.
Key words: JMML; myelodysplastic syndrome; transplantation
INTRODUCTION Allogeneic hematopoietic stem cell transplantation (AHSCT) represents the only curative treatment for the majority of pediatric Myelodysplastic Syndromes (MDS) [1,2]. Pediatric MDS are less frequent than in adults [2] and also different from them in several aspects. The classification, based on a modification of the WHO classification, recognizes the following three major diagnostic groups: juvenile myelomonocytic leukemia (JMML), myeloid leukemia of Down syndrome (ML-DS), and MDS occurring as de novo (primary MDS, pMDS) and secondary to treatment or previous hematologic disease (secondary MDS, sMDS) [3]. MDS are also subdivided into refractory cytopenia of childhood (RCC), refractory anemia with excess blasts (RAEB) and RAEB in transformation (RAEB-T, with 20–30% blasts) [3]. According to Luna-Fineman et al., about 30% of pediatric patients with MDS experience transformation to acute leukemia, usually within 2 years from diagnosis [4]. The IPSS (International Prognosis Scoring System) prognosis index, widely validated in adults, has limited value in pediatric patients [5,6]. Moreover, while the most common cytogenetic abnormalities in pediatric MDS are monosomy 7 and other 7 abnormalities, certain favorable karyotypes in adults, such as 5q-, 20q-, and –Y, are rare in pediatric patients [4]. These clinical and biological differences between MDS in children and MDS in adults justify that transplant outcomes should be analyzed separately. AHSCT is indicated in JMML, RCC with monosomy 7, RCC with severe cytopenia, and advanced MDS [2]. In this study, we retrospectively analyze the outcomes of AHSCT in pediatric patients with MDS.
METHODS A retrospective analysis of 52 consecutive patients under the age of 20 with a diagnosis of pMDS, sMDS, and JMML was performed. Patients underwent AHSCT in six centers (two patients underwent two AHSCT) between 1996 and 2013. These six institutions C
contribute to 80% of all the pediatric transplants performed in Argentina. We followed the classification scheme proposed by Hasle et al. to define minimal criteria for MDS, and diagnostic criteria for JMML, pMDS, and sMDS [3]. Criteria for AHSCT were not formalized since no protocol has been established in this setting in Argentina. Conditioning regimen, graft-versus-host disease (GVHD) prophylaxis and supportive care were accomplished according to each center’s protocol. Patients and/or their respective legal guardians signed an informed consent form. The Independent Ethics Committee for Health Research from Hospital Privado de Cordoba approved the study. Overall survival (OS) was defined as the time from AHSCT to death or last follow-up. Disease-free survival (DFS) was defined as the time from AHSCT to any event (relapse of MDS or death from any cause) or last follow-up. DFS and OS were calculated using the Kaplan–Meier estimator. The log–rank test was used for comparisons of Kaplan–Meier curves. A Cox’s proportional hazard model was used to determine the significance of multiple variables in determining these outcomes. Factors with at least borderline statistical significance in the univariate analysis (P < 0.1) were included in the model. Relapse and non-relapse mortality (NRM) rates were estimated using cumulative incidence analysis and 1
Hospital Privado Centro Me´dico de Cordoba, Cordoba, Argentina; Instituto Universitario de Ciencias Biome´dicas de Cordoba (IUCBC), Cordoba, Argentina; 3Hospital de Pediatrı´a SAMIC Prof. Dr. Juan Pedro Garrahan, Buenos Aires, Argentina; 4Hospital Austral, Pilar, Argentina; 5Hospital de Ni~nos Sor Marı´a Ludovica de La Plata, Buenos Aires, Argentina; 6Hospital Italiano de La Plata, Buenos Aires, Argentina; 7CEHT, Fundacion Favaloro, Buenos Aires, Argentina 2
Conflict of interest: Nothing to declare.
Correspondence to: Ana L. Basquiera, Hospital Privado Centro Me´dico de Cordoba, Naciones Unidas 346 (5016), Cordoba. E-mail: [email protected] Received 15 April 2014; Accepted 5 August 2014
2014 Wiley Periodicals, Inc. DOI 10.1002/pbc.25238 Published online 27 September 2014 in Wiley Online Library (wileyonlinelibrary.com).
154
Basquiera et al.
considered as competing risks. The difference between cumulative incidence curves in the presence of a competing risk was tested using the Gray method [7]. Fisher’s exact test was used to compare categorical variables and Mann–Whitney U test for continuous variables. Two-tailed P values < 0.05 were considered statistically significant. Analyses were performed with StatsDirect v2, 7, 9 (StatsDirect Ltd., Altrincham, UK) and EZR v1.24 (Saitama, Japan), which is a graphical user interface for R (The R Foundation for Statistical Computing) [8].
RESULTS General Data Patients’ characteristics are shown in Table I. At diagnosis, 17 patients had advanced MDS (RAEB and RAEB-T). At transplantation, from the 26 cases of RCC, two (4%) progressed to RAEB and one to acute myeloid leukemia (AML); 6 (35%) out of 17 cases of RAEB/RAEB-T progressed to AML. In one patient with RAEB, MDS was secondary to Kostmann syndrome. Seven patients with advanced MDS received AML-type chemotherapy prior to transplantation (two patients achieved complete remission), and two patients received hypomethylating agents. Mean blasts at transplantation (n ¼ 18) were 10% (range: 0–32). Five transplants were performed between 1996 and 2000, 12 between 2001 and 2005, 12 between 2006 and 2010, and 25 between 2011 and 2013. Mean age at transplantation was 9 years old (range: 2–19). Donors were related in 32 cases (59%) and unrelated in 22 cases (41%; 9 HLA-mismatched). Number of unrelated donors had a progressive temporal rise from 20% in the first period to 56% in the last period included. Median time from diagnosis to transplantation was 12.6 months (range: 1–45); in related donors, it was 6 months (range: 1– 36) and in unrelated donors, it was 21 months (range: 4.3–45) (P < 0.0001). The hematopoietic stem cell sources were as follows: TABLE I. Patients’ Characteristics (n ¼ 52) Variable Age at diagnosis, years, mean (range) Sex, no., male/female Time from diagnosis to transplantation, months, median (range) MDS type at diagnosis, n˚. (%) Refractory cytopenia Refractory anemia with excess blasts Refractory anemia with excess blasts in transformation Juvenile myelomonocytic leukemia Cytogenetics at diagnosis, n˚. (%) Normal Chromosome 7 abnormalities Trisomy 8 Chromosome 11 deletion Chromosome 5 deletion Missing CMV Serology, n˚. (%) Positive Negative Missing
8 (1–16) 35/17 12.6 (1–45)
MDS, myelodysplastic syndrome; CMV, cytomegalovirus. Pediatr Blood Cancer DOI 10.1002/pbc
26 (50) 9 (18) 8 (15) 9 (17) 25 11 3 1 1 11
(48) (21) (6) (2) (2) (21)
45 (87) 5 (9) 2 (4)
bone marrow, 34 cases (63%); peripheral blood, 14 cases (26%); and umbilical cord blood (UCB), 6 cases (11%; 5 unrelated).
Conditioning Regimen A myeloablative conditioning regimen was utilized for all patients. Regimens were: BuCy200 (n: 25; 45%); BuCyMel or BuCyVp16, (n: 14; 25%); BuCy120, (n: 8; 15%); FluBu4, (n: 5; 9%); and TBI þ combinations, (n: 2; 4%). Two patients underwent a second transplant from the same donor due to disease relapse: one patient with RCC and conditioned with BuCy200 received BuCy200 þ antithymocyte globulin 12 months after the first transplant; and another patient with JMML and conditioned with BuCyMel received FLUTBI-Thiotepa 15 months after the first transplant [9]. Prophylaxis against acute GVHD consisted of cyclosporine in 38 cases and tacrolimus in 15 cases; one patient who underwent a syngeneic transplant did not receive prophylaxis. More patients on tacrolimus underwent transplant with an unrelated donor (10/15 vs 12/26; P ¼ 0,026), and all of them were after 2009 [10]. In unrelated donor cases (22 cases, 41%), anti-thymocyte globulin was added to the conditioning regimen. Only one patient who had undergone UCB transplantation presented with primary graft failure.
Graft-Versus-Host Disease Twenty-four (44%) cases developed acute graft versus-host disease (GVHD) of different degrees of severity, 10 (18%) of them were grade III-IV. There was no higher risk of acute GVHD when using unrelated donors (RR 1.7; CI 95% 0.9–3.3; P ¼ 0.132), peripheral blood as a source of hematopoietic stem cells (RR 1.67; CI 95% 0.7–4.1: P ¼ 0.352) or cyclosporine based prophylaxis (RR 0.88; CI 95% 0.6–1.2; P ¼ 0.479). Patients who survived beyond day 100 developed chronic GVHD in 14 out of 41 cases (34%), eight (19%) of whom presented with extensive involvement. Ten patients had previously presented with acute GVHD. The use of unrelated donors did not increase the risk of chronic GVHD (RR 1.6; CI 95% 0.7–3.8; P ¼ 0.298) and there was a non-significant trend towards increased chronic GVHD when peripheral blood was used as a source of hematopoietic stem cells (RR 2.2; CI 95% 0.9–5.3; P ¼ 0.089).
Relapse and Non-Relapse Mortality Eleven patients relapsed at a median time of 9 months (range: 2–15) post-transplantation. The cumulative incidence of relapse was of 21% globally; and of 2% at 30 days, 11% at 90 days and 21% at a year after transplantation. Two patients were rescued with a second transplant; at follow-up they were alive and disease-free. Factors associated with a higher incidence of relapse were conditioning regimens different to BuCy200 and the absence of acute GVHD (Table II). The cumulative incidence of NRM was of 27% globally; and of 6% at 30 days, 18% at 90 days and 25% at a year after transplantation. The causes of the 14 non-relapse deaths were: GVHD, seven cases (50%); infection without GVHD, four cases (28%); hepatic veno-occlusive disease, one case; idiopathic pneumonia, one case; and other causes, one case. The analysis of factors associated with relapse and NRM is shown in Table II. Age equal or older than 9 years (Fig. 1) and GVHD development were associated with higher NRM. Non-difference in NRM was found along time; however, in line with the increasing number of unrelated donors, a concomitant non-significant rise in the median time from diagnosis to transplant (5 months, 7.5 months,
Transplantation in Pediatric Myelodysplasia
155
TABLE II. Relapse and Non-relapse Mortality Cumulative Incidence Analysis Factors Age at transplantation 4 months Donor type Related Unrelated Source Bone marrow Peripheral blood Umbilical cord blood Transplant year 1996–2000 2001–2005 2006–2010 2011–2013 Conditioning BuCy200 Others Acute GVHD Yes No Chronic GVHDa Yes No
N
Relapse at 1 year (95% CI)
28 26
26% (11–43%) 17% (5–34%)
35 19
24% (11–39%) 17% (4–37%)
10 44
31% (6–62%) 19% (10–32%)
14 39
24% (5–49%) 21% (10–35%)
32 22
20% (8–37%) 23% (8–43%)
34 14 6
13% (5–32%) 30% (8–56%) 50% (7–83%)
5 12 12 25
20% 17% 27% 21%
24 30
5% (5–47%) 35% (19–53%)
25 29
9% (2–26%) 31% (15–49%)
14 27
8% (1–30%) 31% (14–69%)
P-value
NRM at 1 year (95% CI)
0.412
P-value 0.034
16% (5–32%) 41% (21–61%) 0.467
0.611 28% (13–44%) 21% (6–42%)
0.393
0.197 11% (1–42%) 25% (16–42%)
0.99
0.059 7% (1–31%) 29% (5–43%)
0.631
0.129 17% (6–32%) 32% (14–52%)
0.058
0.514 30% (18–50%) 20% (2–51%) 33% (2–74%)
0.969 (4–62%) (4–37%) (8–51%) (7–39%)
0.714 20% 33% 17% 26%
(4–63%) (9–60%) (2–43%) (10–45%)
0.007
0.784 25% (10–44%) 25% (11–43%)
0.042
0.005 39% (19–58%) 10% (3–29%)
0.107
16% (2–42%) 0%
0.003
NRM, non-relapse mortality; MDS, myelodysplastic syndrome; CI, confidence interval; JMML, juvenile myelomonocytic leukemia; GVHD, graft-versus-host disease. aPatients who survive day þ100.
10 months, and 16 months; P ¼ 0.245), and an increasing proportion of HLA-mismatched unrelated donors (0%, 0%, 8%, 32%; P ¼ 0.042) were detected.
Survival
Fig. 1. Cumulative incidence of non-relapse mortality (NRM) according to age at transplantation (P ¼ 0.034). Pediatr Blood Cancer DOI 10.1002/pbc
With a median follow-up of 22.1 months (range: 6.5–187 months), 5 year DFS and 5-year OS for all patients were of 50% (95% CI 35–63%) and of 55% (95% CI 40–67%) respectively; for RCC 64% (95% CI 44–84%) and 68% (95% CI 40–80%) respectively; for JMML 57% (95% CI 22–81%) and 67% (95% CI 29–88%) respectively; and for advanced MDS at transplant, 30% (95% CI 12–50%) and 33% (95% CI 13–54%) respectively (Fig. 2). The univariate analysis identified age at transplantation, conditioning regimen, and source of hematopoietic stem cells as variables affecting the OS; and the donor type, conditioning regimen and source of hematopoietic stem cells as significant variables for DFS (Table III). In the multivariate analysis, the use of umbilical cord (HR 4.07; P ¼ 0.025) and age equal or older than 9 years at the moment of AHSCT (HR 3.28; P ¼ 0.017) were associated with lower OS. Variables associated with DFS were not confirmed in the multivariate analysis (Table IV).
156
Basquiera et al.
Fig. 2. Overall survival for patients with refractory cytopenia, juvenile myelomonocytic leukemia (JMML) and advanced myelodysplasia at transplant (P ¼ 0.035).
DISCUSSION We present a series of patients under the age of 20 years who underwent AHSCT for MDS in six Argentinean centers. It has not been reported another study with such an extensive series of pediatric patients with MDS who underwent AHSCT in Latin America. In 2004, the Brazilian group of pediatric MDS published a series of 173 patients with pediatric MDS, from which only 11 underwent AHSCT [11].
MDS are infrequent diseases in pediatric patients with a dismal outcome when treated only with conventional chemotherapy [1]. The use of unrelated donor has allowed gaining access to more patients for transplants. In our series, 41% of the cases underwent AHSCT from an unrelated donor and this factor did not affect the NRM nor the OS, the same as it has been previously published [6,12,13]. Our most important finding was an OS of 55% at 5 years. This percentage compares favorably to what has been reported by other studies, with a range of 50–63% [6,13–17]. In patients with JMML, the most important series reported an OS of 64% at 5 years [12]; in our series, the survival of patients with JMML was of 67% at 5 years. These data demonstrate a clear advantage on OS when it is compared to historic series of 33% for patients with JMML who did not undergo AHSCT [4]. Previous studies of transplantation in pediatric MDS have shown that approximately 20% of the patients relapse whereas 21–35% of them will experience NRM [12–15,17]. Our results show a relapse (21%) and NRM (27%) rate similar to the figures reported [12– 15,17]. The use of BuCy200 and the presence of acute GVHD (all grades) were associated with lower relapse rate; however, these variables were not confirmed in the multivariate analysis for DFS. NRM was clearly related to the age at the moment of transplantation and the development of GVHD. GVHD was the main cause of NRM in our series. Several authors pointed out that occurrence of GVHD was associated with NRM in pediatric patients with MDS [13,18]. In the multivariate analysis for OS, age equal or older than 9 years was an independent factor for worse OS. Even though with a discrete difference in the cut-off point, this finding has also been previously described [13,15,19], and genetic or environmental variations could be the underlying cause of this difference. Cytomegalovirus (CMV) seropositivity rate (88%) in children from our series is higher than European [12] or American [14] series
TABLE III. Univariate Analysis for Overall Survival and Disease-Free Survival Variables Age at transplantation
Allogeneic Hematopoietic Stem Cell Transplantation in Pediatric Myelodysplastic Syndromes: A Multicenter Experience From Argentina Ana L. Basquiera, MD,1,2* Silvia Pizzi, MD,3 Agustı´n Gonza´lez Correas, MD,3 Pablo G. Longo, MD,4 Wanda C. Goldman, MD,5 Marı´a V. Prates, MD,6 Sandra Formisano, MD,5 Gustavo Kusminisky, MD,4 Leonardo Feldman, MD,7 Adriana R. Berretta, MD,1 Juan J. Garcı´a, MD,1,2 and Raquel Staciuk, MD3 Background. Allogeneic hematopoietic stem cell transplantation (AHSCT) represents the only curative treatment for the majority of pediatric patients with Myelodysplastic Syndrome (MDS). We aimed to evaluate overall survival (OS), disease-free survival (DFS), nonrelapse mortality (NRM) and relapse incidence in children who underwent AHSCT for MDS in six institutions from Argentina. Procedure. A retrospective analysis of 54 AHSCT was carried out in 52 patients (mean age: 9 years; range: 2–19; 35 males). Results. MDS subtypes were refractory cytopenia of childhood (RCC) (n: 26, 50%), refractory anemia with excess blasts (RAEB) (n: 9, 18%), RAEB in transformation (RAEB-T) (n: 8, 15%) and juvenile myelomonocytic leukemia (JMML) (n: 9, 17%). At time of transplant, seven (13%) patients transformed to acute myeloid leukemia (AML) and two
patients with RCC to RAEB. Donors were related in 32 cases (59%) and the stem cells source was: bone marrow (63%), peripheral blood (26%), and umbilical cord blood (11%). Five-year DFS and OS were 50% and 55% respectively; and for patients with JMML, 57% and 67% respectively. Cumulative incidence of NRM and relapse were 27% and 21% respectively. In the multivariate analysis, umbilical cord blood (HR 4.07; P ¼ 0.025) and age 9 years at transplantation (HR 3.28; P ¼ 0.017) were associated with lower OS; age and graftversus-host disease (GVHD) had a higher NRM. Conclusions. In our series, more than half of the patients achieved long term OS with AHSCT. Less toxic conditioning regimens or more intensive GVHD prophylaxis could lead to better results in some children. Pediatr Blood Cancer 2015;62:153–157. # 2014 Wiley Periodicals, Inc.
Key words: JMML; myelodysplastic syndrome; transplantation
INTRODUCTION Allogeneic hematopoietic stem cell transplantation (AHSCT) represents the only curative treatment for the majority of pediatric Myelodysplastic Syndromes (MDS) [1,2]. Pediatric MDS are less frequent than in adults [2] and also different from them in several aspects. The classification, based on a modification of the WHO classification, recognizes the following three major diagnostic groups: juvenile myelomonocytic leukemia (JMML), myeloid leukemia of Down syndrome (ML-DS), and MDS occurring as de novo (primary MDS, pMDS) and secondary to treatment or previous hematologic disease (secondary MDS, sMDS) [3]. MDS are also subdivided into refractory cytopenia of childhood (RCC), refractory anemia with excess blasts (RAEB) and RAEB in transformation (RAEB-T, with 20–30% blasts) [3]. According to Luna-Fineman et al., about 30% of pediatric patients with MDS experience transformation to acute leukemia, usually within 2 years from diagnosis [4]. The IPSS (International Prognosis Scoring System) prognosis index, widely validated in adults, has limited value in pediatric patients [5,6]. Moreover, while the most common cytogenetic abnormalities in pediatric MDS are monosomy 7 and other 7 abnormalities, certain favorable karyotypes in adults, such as 5q-, 20q-, and –Y, are rare in pediatric patients [4]. These clinical and biological differences between MDS in children and MDS in adults justify that transplant outcomes should be analyzed separately. AHSCT is indicated in JMML, RCC with monosomy 7, RCC with severe cytopenia, and advanced MDS [2]. In this study, we retrospectively analyze the outcomes of AHSCT in pediatric patients with MDS.
METHODS A retrospective analysis of 52 consecutive patients under the age of 20 with a diagnosis of pMDS, sMDS, and JMML was performed. Patients underwent AHSCT in six centers (two patients underwent two AHSCT) between 1996 and 2013. These six institutions C
contribute to 80% of all the pediatric transplants performed in Argentina. We followed the classification scheme proposed by Hasle et al. to define minimal criteria for MDS, and diagnostic criteria for JMML, pMDS, and sMDS [3]. Criteria for AHSCT were not formalized since no protocol has been established in this setting in Argentina. Conditioning regimen, graft-versus-host disease (GVHD) prophylaxis and supportive care were accomplished according to each center’s protocol. Patients and/or their respective legal guardians signed an informed consent form. The Independent Ethics Committee for Health Research from Hospital Privado de Cordoba approved the study. Overall survival (OS) was defined as the time from AHSCT to death or last follow-up. Disease-free survival (DFS) was defined as the time from AHSCT to any event (relapse of MDS or death from any cause) or last follow-up. DFS and OS were calculated using the Kaplan–Meier estimator. The log–rank test was used for comparisons of Kaplan–Meier curves. A Cox’s proportional hazard model was used to determine the significance of multiple variables in determining these outcomes. Factors with at least borderline statistical significance in the univariate analysis (P < 0.1) were included in the model. Relapse and non-relapse mortality (NRM) rates were estimated using cumulative incidence analysis and 1
Hospital Privado Centro Me´dico de Cordoba, Cordoba, Argentina; Instituto Universitario de Ciencias Biome´dicas de Cordoba (IUCBC), Cordoba, Argentina; 3Hospital de Pediatrı´a SAMIC Prof. Dr. Juan Pedro Garrahan, Buenos Aires, Argentina; 4Hospital Austral, Pilar, Argentina; 5Hospital de Ni~nos Sor Marı´a Ludovica de La Plata, Buenos Aires, Argentina; 6Hospital Italiano de La Plata, Buenos Aires, Argentina; 7CEHT, Fundacion Favaloro, Buenos Aires, Argentina 2
Conflict of interest: Nothing to declare.
Correspondence to: Ana L. Basquiera, Hospital Privado Centro Me´dico de Cordoba, Naciones Unidas 346 (5016), Cordoba. E-mail: [email protected] Received 15 April 2014; Accepted 5 August 2014
2014 Wiley Periodicals, Inc. DOI 10.1002/pbc.25238 Published online 27 September 2014 in Wiley Online Library (wileyonlinelibrary.com).
154
Basquiera et al.
considered as competing risks. The difference between cumulative incidence curves in the presence of a competing risk was tested using the Gray method [7]. Fisher’s exact test was used to compare categorical variables and Mann–Whitney U test for continuous variables. Two-tailed P values < 0.05 were considered statistically significant. Analyses were performed with StatsDirect v2, 7, 9 (StatsDirect Ltd., Altrincham, UK) and EZR v1.24 (Saitama, Japan), which is a graphical user interface for R (The R Foundation for Statistical Computing) [8].
RESULTS General Data Patients’ characteristics are shown in Table I. At diagnosis, 17 patients had advanced MDS (RAEB and RAEB-T). At transplantation, from the 26 cases of RCC, two (4%) progressed to RAEB and one to acute myeloid leukemia (AML); 6 (35%) out of 17 cases of RAEB/RAEB-T progressed to AML. In one patient with RAEB, MDS was secondary to Kostmann syndrome. Seven patients with advanced MDS received AML-type chemotherapy prior to transplantation (two patients achieved complete remission), and two patients received hypomethylating agents. Mean blasts at transplantation (n ¼ 18) were 10% (range: 0–32). Five transplants were performed between 1996 and 2000, 12 between 2001 and 2005, 12 between 2006 and 2010, and 25 between 2011 and 2013. Mean age at transplantation was 9 years old (range: 2–19). Donors were related in 32 cases (59%) and unrelated in 22 cases (41%; 9 HLA-mismatched). Number of unrelated donors had a progressive temporal rise from 20% in the first period to 56% in the last period included. Median time from diagnosis to transplantation was 12.6 months (range: 1–45); in related donors, it was 6 months (range: 1– 36) and in unrelated donors, it was 21 months (range: 4.3–45) (P < 0.0001). The hematopoietic stem cell sources were as follows: TABLE I. Patients’ Characteristics (n ¼ 52) Variable Age at diagnosis, years, mean (range) Sex, no., male/female Time from diagnosis to transplantation, months, median (range) MDS type at diagnosis, n˚. (%) Refractory cytopenia Refractory anemia with excess blasts Refractory anemia with excess blasts in transformation Juvenile myelomonocytic leukemia Cytogenetics at diagnosis, n˚. (%) Normal Chromosome 7 abnormalities Trisomy 8 Chromosome 11 deletion Chromosome 5 deletion Missing CMV Serology, n˚. (%) Positive Negative Missing
8 (1–16) 35/17 12.6 (1–45)
MDS, myelodysplastic syndrome; CMV, cytomegalovirus. Pediatr Blood Cancer DOI 10.1002/pbc
26 (50) 9 (18) 8 (15) 9 (17) 25 11 3 1 1 11
(48) (21) (6) (2) (2) (21)
45 (87) 5 (9) 2 (4)
bone marrow, 34 cases (63%); peripheral blood, 14 cases (26%); and umbilical cord blood (UCB), 6 cases (11%; 5 unrelated).
Conditioning Regimen A myeloablative conditioning regimen was utilized for all patients. Regimens were: BuCy200 (n: 25; 45%); BuCyMel or BuCyVp16, (n: 14; 25%); BuCy120, (n: 8; 15%); FluBu4, (n: 5; 9%); and TBI þ combinations, (n: 2; 4%). Two patients underwent a second transplant from the same donor due to disease relapse: one patient with RCC and conditioned with BuCy200 received BuCy200 þ antithymocyte globulin 12 months after the first transplant; and another patient with JMML and conditioned with BuCyMel received FLUTBI-Thiotepa 15 months after the first transplant [9]. Prophylaxis against acute GVHD consisted of cyclosporine in 38 cases and tacrolimus in 15 cases; one patient who underwent a syngeneic transplant did not receive prophylaxis. More patients on tacrolimus underwent transplant with an unrelated donor (10/15 vs 12/26; P ¼ 0,026), and all of them were after 2009 [10]. In unrelated donor cases (22 cases, 41%), anti-thymocyte globulin was added to the conditioning regimen. Only one patient who had undergone UCB transplantation presented with primary graft failure.
Graft-Versus-Host Disease Twenty-four (44%) cases developed acute graft versus-host disease (GVHD) of different degrees of severity, 10 (18%) of them were grade III-IV. There was no higher risk of acute GVHD when using unrelated donors (RR 1.7; CI 95% 0.9–3.3; P ¼ 0.132), peripheral blood as a source of hematopoietic stem cells (RR 1.67; CI 95% 0.7–4.1: P ¼ 0.352) or cyclosporine based prophylaxis (RR 0.88; CI 95% 0.6–1.2; P ¼ 0.479). Patients who survived beyond day 100 developed chronic GVHD in 14 out of 41 cases (34%), eight (19%) of whom presented with extensive involvement. Ten patients had previously presented with acute GVHD. The use of unrelated donors did not increase the risk of chronic GVHD (RR 1.6; CI 95% 0.7–3.8; P ¼ 0.298) and there was a non-significant trend towards increased chronic GVHD when peripheral blood was used as a source of hematopoietic stem cells (RR 2.2; CI 95% 0.9–5.3; P ¼ 0.089).
Relapse and Non-Relapse Mortality Eleven patients relapsed at a median time of 9 months (range: 2–15) post-transplantation. The cumulative incidence of relapse was of 21% globally; and of 2% at 30 days, 11% at 90 days and 21% at a year after transplantation. Two patients were rescued with a second transplant; at follow-up they were alive and disease-free. Factors associated with a higher incidence of relapse were conditioning regimens different to BuCy200 and the absence of acute GVHD (Table II). The cumulative incidence of NRM was of 27% globally; and of 6% at 30 days, 18% at 90 days and 25% at a year after transplantation. The causes of the 14 non-relapse deaths were: GVHD, seven cases (50%); infection without GVHD, four cases (28%); hepatic veno-occlusive disease, one case; idiopathic pneumonia, one case; and other causes, one case. The analysis of factors associated with relapse and NRM is shown in Table II. Age equal or older than 9 years (Fig. 1) and GVHD development were associated with higher NRM. Non-difference in NRM was found along time; however, in line with the increasing number of unrelated donors, a concomitant non-significant rise in the median time from diagnosis to transplant (5 months, 7.5 months,
Transplantation in Pediatric Myelodysplasia
155
TABLE II. Relapse and Non-relapse Mortality Cumulative Incidence Analysis Factors Age at transplantation 4 months Donor type Related Unrelated Source Bone marrow Peripheral blood Umbilical cord blood Transplant year 1996–2000 2001–2005 2006–2010 2011–2013 Conditioning BuCy200 Others Acute GVHD Yes No Chronic GVHDa Yes No
N
Relapse at 1 year (95% CI)
28 26
26% (11–43%) 17% (5–34%)
35 19
24% (11–39%) 17% (4–37%)
10 44
31% (6–62%) 19% (10–32%)
14 39
24% (5–49%) 21% (10–35%)
32 22
20% (8–37%) 23% (8–43%)
34 14 6
13% (5–32%) 30% (8–56%) 50% (7–83%)
5 12 12 25
20% 17% 27% 21%
24 30
5% (5–47%) 35% (19–53%)
25 29
9% (2–26%) 31% (15–49%)
14 27
8% (1–30%) 31% (14–69%)
P-value
NRM at 1 year (95% CI)
0.412
P-value 0.034
16% (5–32%) 41% (21–61%) 0.467
0.611 28% (13–44%) 21% (6–42%)
0.393
0.197 11% (1–42%) 25% (16–42%)
0.99
0.059 7% (1–31%) 29% (5–43%)
0.631
0.129 17% (6–32%) 32% (14–52%)
0.058
0.514 30% (18–50%) 20% (2–51%) 33% (2–74%)
0.969 (4–62%) (4–37%) (8–51%) (7–39%)
0.714 20% 33% 17% 26%
(4–63%) (9–60%) (2–43%) (10–45%)
0.007
0.784 25% (10–44%) 25% (11–43%)
0.042
0.005 39% (19–58%) 10% (3–29%)
0.107
16% (2–42%) 0%
0.003
NRM, non-relapse mortality; MDS, myelodysplastic syndrome; CI, confidence interval; JMML, juvenile myelomonocytic leukemia; GVHD, graft-versus-host disease. aPatients who survive day þ100.
10 months, and 16 months; P ¼ 0.245), and an increasing proportion of HLA-mismatched unrelated donors (0%, 0%, 8%, 32%; P ¼ 0.042) were detected.
Survival
Fig. 1. Cumulative incidence of non-relapse mortality (NRM) according to age at transplantation (P ¼ 0.034). Pediatr Blood Cancer DOI 10.1002/pbc
With a median follow-up of 22.1 months (range: 6.5–187 months), 5 year DFS and 5-year OS for all patients were of 50% (95% CI 35–63%) and of 55% (95% CI 40–67%) respectively; for RCC 64% (95% CI 44–84%) and 68% (95% CI 40–80%) respectively; for JMML 57% (95% CI 22–81%) and 67% (95% CI 29–88%) respectively; and for advanced MDS at transplant, 30% (95% CI 12–50%) and 33% (95% CI 13–54%) respectively (Fig. 2). The univariate analysis identified age at transplantation, conditioning regimen, and source of hematopoietic stem cells as variables affecting the OS; and the donor type, conditioning regimen and source of hematopoietic stem cells as significant variables for DFS (Table III). In the multivariate analysis, the use of umbilical cord (HR 4.07; P ¼ 0.025) and age equal or older than 9 years at the moment of AHSCT (HR 3.28; P ¼ 0.017) were associated with lower OS. Variables associated with DFS were not confirmed in the multivariate analysis (Table IV).
156
Basquiera et al.
Fig. 2. Overall survival for patients with refractory cytopenia, juvenile myelomonocytic leukemia (JMML) and advanced myelodysplasia at transplant (P ¼ 0.035).
DISCUSSION We present a series of patients under the age of 20 years who underwent AHSCT for MDS in six Argentinean centers. It has not been reported another study with such an extensive series of pediatric patients with MDS who underwent AHSCT in Latin America. In 2004, the Brazilian group of pediatric MDS published a series of 173 patients with pediatric MDS, from which only 11 underwent AHSCT [11].
MDS are infrequent diseases in pediatric patients with a dismal outcome when treated only with conventional chemotherapy [1]. The use of unrelated donor has allowed gaining access to more patients for transplants. In our series, 41% of the cases underwent AHSCT from an unrelated donor and this factor did not affect the NRM nor the OS, the same as it has been previously published [6,12,13]. Our most important finding was an OS of 55% at 5 years. This percentage compares favorably to what has been reported by other studies, with a range of 50–63% [6,13–17]. In patients with JMML, the most important series reported an OS of 64% at 5 years [12]; in our series, the survival of patients with JMML was of 67% at 5 years. These data demonstrate a clear advantage on OS when it is compared to historic series of 33% for patients with JMML who did not undergo AHSCT [4]. Previous studies of transplantation in pediatric MDS have shown that approximately 20% of the patients relapse whereas 21–35% of them will experience NRM [12–15,17]. Our results show a relapse (21%) and NRM (27%) rate similar to the figures reported [12– 15,17]. The use of BuCy200 and the presence of acute GVHD (all grades) were associated with lower relapse rate; however, these variables were not confirmed in the multivariate analysis for DFS. NRM was clearly related to the age at the moment of transplantation and the development of GVHD. GVHD was the main cause of NRM in our series. Several authors pointed out that occurrence of GVHD was associated with NRM in pediatric patients with MDS [13,18]. In the multivariate analysis for OS, age equal or older than 9 years was an independent factor for worse OS. Even though with a discrete difference in the cut-off point, this finding has also been previously described [13,15,19], and genetic or environmental variations could be the underlying cause of this difference. Cytomegalovirus (CMV) seropositivity rate (88%) in children from our series is higher than European [12] or American [14] series
TABLE III. Univariate Analysis for Overall Survival and Disease-Free Survival Variables Age at transplantation
