¹Division of Hematology/Oncology, The Hospital for Sick Children, and the Department of Pediatrics, University of Toronto Faculty of Medicine, Toronto, Ontario, Canada

²Department of Biostatistics and Epidemiology, St Jude Children's Research Hospital, Memphis, Tennessee, USA

Correspondence to: Dr M H Freedman, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada, M5G 1X8

Multi-agent immunosuppressive therapy has produced improved survival for severe acquired aplastic anemia in children. Recently, some investigators have suggested that immunosuppressive therapy may replace bone marrow transplantation as first-line therapy for this disorder. To assess its validity, we compared the outcomes of bone marrow transplantation vsimmunosuppressive therapy in one institution from 1987 to 1997. We studied 46 consecutive patients less than 18 years of age who presented between January 1987 and April 1997. Inherited marrow failure syndromes and myelodysplastic syndromes were excluded. Patients received immunosuppressive therapy vs bone marrow transplantation based on availability of HLA-matched donors. The main outcome measures were survival, complete marrow and hematological remission, or partial remission but achieving independence from transfusional support. Twenty patients received multi-agent immunosuppressive therapy (cyclosporine, antithymocyte globulin and methylprednisolone); 11 attained complete remission and three partial remission for a transfusion-independent survival of 70%. Six patients died of infectious and hemorrhagic complications. Twenty-six patients were transplanted and 24 (93%) achieved complete remission; one achieved a PR, 25 remain transfusion independent with a median follow-up of 5.9 years or 70 months. One patient developed AML 34 months after successful transplant and one patient died due to graft failure and complications of transplant. There has been a striking improvement in survival for pediatric patients treated with multi-agent immunosuppression in the last decade. However, transplantation results have also improved and this remains the definitive first-line therapy for severe acquired aplastic anemia in this age group. Bone Marrow Transplantation (2000) 26, 1149-1156.

aplastic anemia; immunosuppressive therapy; bone marrow transplantation

Aplastic anemia is defined as peripheral blood pancytopenia, bone marrow hypocellularity with fatty replacement, and the absence of an underlying hematopoietic malignancy or a myelodysplastic disorder. When there is no evidence of a defined syndrome or familial tendency, the aplastic anemia is considered to be acquired. The designation, 'severe aplastic anemia' applies when there is moderate (25-50%) or marked (<25%) reduction in marrow cellularity demonstrated on biopsy in combination with two of the following findings in the peripheral blood: absolute neutrophil count <0.5 ´ 10⁹/l; platelet count <20.0 ´ 10⁹/l; reticulocyte count <1% or <20 ´ 10⁹/l after correction for hematocrit.¹

Prior to 1970, severe aplastic anemia was managed by supportive care only and the 2-year survival was 20% or less.² In 1972, Thomas et al³ reported a therapeutic breakthrough by demonstrating the curative potential of bone marrow transplantation using marrow cells from healthy sibling donors. As experience with transplantation broadened,⁴ it became clear that the procedure could be recommended as frontline therapy for severe aplastic anemia, especially in children, if a matched donor was available.

Immunosuppressive therapy for aplastic anemia developed concurrently in the same era following the seminal report of Mathé et al⁵ on the ameliorating effects of antilymphocyte globulin. Subsequent experience showed improved survival rates with antilymphocyte globulin used as a single agent⁶ or combined with other immunosuppressives.^2,7,8 This advance provided options for treatment of severe aplastic anemia, especially in patients lacking a matched donor for transplantation.

By 1990, several retrospective analyses suggested a survival advantage of marrow transplantation over immunosuppressive therapy in pediatric cases of severe aplastic anemia.^9,10,11 Very recently, with the advent of cyclosporine-based immunosuppression protocols and improved supportive measures, equivalent survival rates have been reported.^12,13 These studies have formed the basis for the recent suggestion that immunosuppressive therapy should be used as initial treatment for all children with severe aplastic anemia regardless of donor availability for transplantation.

We have previously published the outcome of a similar series of patients with severe aplastic anemia treated between 1977 and 1987, which clearly demonstrates the advantage of bone marrow transplantation over immunosuppressive therapy. To determine whether the improvement in immunosuppressive regimens used in the past decade has changed outcomes substantially, we critically reviewed the outcome, and survival of all childhood cases of severe aplastic anemia in our institution with respect to treatment given from 1987 to 1997. The power of our study stems from the large number of patients diagnosed and treated over the 10-year period in one center using similar protocols by the same group of experienced pediatric hematologists. Furthermore, we had good historical comparisons from the same institution as we compared our results over the past decade with our previously published results.

Patients and methods

Patient eligibility

Forty-six consecutive patients less than 18 years of age who presented with a diagnosis of severe aplastic anemia between January 1987 and April 1997 were included in the analysis. Patients with constitutional aplastic anemia and myelodysplastic syndromes were excluded. Information obtained from their medical records included age, sex, family history, drug and environmental history, initial signs and symptoms, initial complete blood counts, bone marrow aspiration and biopsy results, type of treatment including type and dates of transfusions, medications and dosages, the marrow transplantation conditioning regimen, and complications, outcomes, and survivals.

Definition

Severe aplastic anemia was defined as moderate (25-50%) or marked (<25%) reduction in marrow cellularity demonstrated on biopsy in combination with two of the following findings in the peripheral blood: absolute neutrophil count <0.5 ´ 10⁹/l; platelet count <20.0 ´ 10⁹/l; reticulocyte count <1% or <20 ´ 10⁹/l after correction for hematocrit.¹

Treatment regimens

Throughout the 10-year study period transplantation was offered as first-line treatment when a fully histocompatible related donor was available. When no suitable donor was available, immunosuppressive therapy was used while a search for a matched unrelated donor was initiated. If transfusion independence was not achieved after 6 months, despite repeat therapy, a matched unrelated bone marrow transplant was performed if a donor was identified.

Immunosuppressive therapy

Twenty patients were treated with immunosuppressive agents. Nineteen were treated with antithymocyte globulin (ATGAM, Pharmacia and Upjohn, Don Mills, ON, Canada), cyclosporine (Sandimmune, Novartis Pharmaceuticals Canada, Dorval, PQ, Canada) and some form of corticosteroid. One patient did not receive cyclosporine. In addition to this treatment, one patient received GM-CSF, two patients received G-CSF, one received IL-3, and two received intravenous gammaglobulin. The total dose of ATGAM ranged from 140 to 160 mg/kg; high-dose methylprednisolone was used concomitantly at doses ranging from 1 to 30 mg/kg for 4-14 days and was then tapered; and cyclosporine was used to maintain a serum trough level of 150-200 g/l until a complete or partial response was achieved and was then tapered and discontinued after at least 3 months of therapy.

Bone marrow transplantation

Twenty-six patients underwent bone marrow transplantation. Twenty-one underwent a matched related bone marrow transplant (BMT-MRD), and five underwent a matched unrelated bone marrow transplant (BMT-MUD). The conditioning regimen in the BMT-MRD group included cyclophosphamide and total body irradiation (TBI) in 12 patients, cyclophosphamide and ATGAM in three patients, cyclophosphamide, rabbit antithymocyte serum (RATS), procarbazine in two patients, cyclophosphamide, ATGAM, procarbazine in two patients, busulfan, cyclophosphamide in one patient and cyclophosphamide alone in five patients. In the one patient who relapsed and required a second transplant, the conditioning regimen initially consisted of only cyclophosphamide but included both cyclophosphamide and TBI for the second transplant. To prevent the occurrence of graft-versus-host disease (GVHD), the prophylaxis regimen consisted of methotrexate and cyclosporine in 17 patients, cyclosporine and prednisone in three patients and cyclosporine alone in six patients. Twenty-one patients had a fully histocompatible related donor. Five, with no matched related donors, who had failed immunosuppressive therapy, were transplanted using marrow from matched unrelated donors. Conditioning in this group consisted of cyclophosphamide and TBI.

Response criteria

Complete response was defined as hemoglobin normal for age, a neutrophil count >1.5 ´ 10⁹/l and a platelet count of >100 ´ 10⁹ g/l. Partial response was defined as transfusion independence with a hemoglobin >80 g/l, neutrophil count >0.5 ´ 10⁹/l and a platelet count of greater than 20 ´ 10⁹/l.¹⁴

Statistical analysis

The method of Kaplan and Meier¹⁵ was used to estimate survival distributions. Associated standard errors were calculated as suggested by Peto et al.¹⁶ Differences in survival between the transplant and IST groups were assessed by the exact log rank test.¹⁷ Cumulative incidence functions of time to complete response were estimated as described in Kalbfleisch and Prentice.¹⁸ Partial response and death were considered competing risks. Differences in the cumulative incidence functions among the treatment groups were compared using Gray's test.¹⁹ The prognostic implications of age at diagnosis (<9 vs 9), gender, months from diagnosis to treatment (<1 vs 1), ANC at diagnosis (<0.2 ´ 10⁹/l vs 0.2 ´ 10⁹/l), and platelet count at diagnosis (<8 ´ 10⁹/l vs 8 ´ 10⁹/l) within the IST group were assessed with the exact log rank statistic.¹⁷

Results

Table 1 summarizes the demographics and clinical characteristics of all patients. Forty-six patients with a median age of 9.2 years were treated. The median age was similar between the two main treatment groups (IST: 9.8 years; BMT-MRD: 8.8 years). There was almost a two to one male preponderance (M:F = 30:16). Causes of the aplasia were idiopathic in 39 and hepatitis in seven.

Results of bone marrow transplantation

Twenty-one patients received an allogeneic bone marrow transplant from a fully histocompatible related donor. One patient had previously failed IST at another institution. The time from initiation of IST to BMT-MRD, in this patient, was 4 months. Twenty (95%) achieved a complete remission (Table 2). One patient, included among the 20, initially achieved a complete remission, experienced graft failure 8 months post transplant and went on to a second transplant and complete remission with a follow-up of 4.3 years (3.3-6.7 years). One patient developed AML 34 months after the successful transplant. One patient died due to graft failure and complications of the transplant (Table 2).

Five patients were referred to the institution to undergo matched unrelated donor transplant. Their previous immunosuppressive therapy regimens were not available. The median time from initiation of IST to MUD-BMT was 4.5 months (range: 2.8-5 months). Within this group, four achieved a complete remission and one a partial remission (Table 2).

The time to response was calculated in the three groups, and the cumulative incidence of time to CR is shown in Figure 1. There is a significant difference in the cumulative incidence of time to CR (P = 0.0003). At 2 months, the cumulative incidence of CR is 20 ± 9% for the IST group, 71 ± 10% for the BMT group, and 60 ± 26% for the BMT-MUD group.

Complications of bone marrow transplantation in the BMT-MRD group included grade I acute GVHD in five, grade II in two, and grade III in one patient. Chronic GVHD was documented in four patients. One patient developed mild chronic GVHD of the skin. One patient developed mild chronic diarrhea. One patient developed moderate chronic GVHD with moderate skin involvement and restrictive lung disease, and mild chronic diarrhea. One patient developed moderate to extensive GVHD with severe skin involvement, chronic diarrhea and hepatitis (Table 3).

Complications of bone marrow transplantation in the BMT-MUD group include grade 1 acute GVHD in two, grade 2 in one, and grade 3 in one patient. Chronic GVHD occurred in four of the five patients with BMT-MUD. Two patients developed mild chronic GVHD: one with mild chronic diarrhea and lichenoid changes of the skin, another with mild scleroderma. Two others developed moderate chronic GVHD manifesting as bronchiolitis obliterans and hepatitis in one, and moderate skin changes accompanied with mild ocular changes and diarrhea (Table 3).

Results of immunosuppressive therapy

Among the 20 patients treated with IST, 11 (55%) have remained in complete remission and three (15%) have remained in partial remission (Table 2) with a median follow-up of 3 years (0.2-6.7 years). Thus, in total, 14 (70%) have remained transfusion-independent. The six other patients died of infectious and hemorrhagic complications (Table 4). Complications included serum sickness secondary to ATGAM in three (15%) patients and hypertension in four (20%). Median time of death after initiation of IST was 2.8 months (range: 0.8-34 months). Other than hypertension, no patient had major complications associated with the use of high-dose methylpredniosolone.

Kaplan-Meier survival curves

The Kaplan-Meier plot of overall survival by treatment is shown in Figure 2. There is no difference between the BMT-MRD group and the BMT MUD group. However, there is a significant difference between the IST group and the combined BMT groups (P = 0.0039). At 3 years, the estimated survival for the IST group is 65 ± 13% compared with 95 ± 5% for the BMT-MRD group and 100% for the BMT-MUD group.

Factors predicting survival within the IST group

We examined factors that may predict survival in those patients treated with IST. Because the number of deaths is small within the group, and the group sizes are small, this analysis should be interpreted with caution. Factors examined were age at diagnosis, months from diagnosis to treatment, ANC at diagnosis, and platelet count at diagnosis. Only ANC at the time of diagnosis had significant predictive ability (P = 0.0058). At 1 year after diagnosis, the estimated survival of those patients with ANC <0.2 ´ 10⁹/l at diagnosis is 25 ± 15% compared with 93 ± 6% for those with ANC >0.2 ´ 10⁹/l.

Discussion

Traditionally, bone marrow transplantation from a fully matched sibling donor has been considered the best available therapy for children with severe aplastic anemia. In several trials, matched sibling donor transplantation for severe aplastic anemia has led to overall survivals varying from 60 to 100%.^{10,20,21,22,23,24,25,26,27,28,29,30,31,32,33} In a recent report by Bacigalupo et al,³⁴ both forms of therapy among 3669 patients treated in Europe between 1976 and 1998 were reviewed. The authors reported that significant progress has been made and the overall risk of failure is low with survival rates of more than 80% for both treatments. Our results are in keeping with these published data regarding improved outcome of transplantation with 95% survival rates in the last decade compared to 79% between 1977 and 1987. This improvement in outcome is due to more effective supportive care for patients and improvement of prophylactic regimens used to minimize graft rejection. The European Bone Marrow Transplantation (EBMT) Registry data reported graft rejection rates of 32% prior to 1980 and 7.6% after 1984.³⁵ Factors responsible for decreasing the incidence of graft rejection include minimizing pretransplant transfusions,^35,36,37,38 use of marrow doses >3 ´ 10⁸ marrow cells/kg,³⁹ and the use of cyclosporine for acute GVHD prophylaxis.^33,35,36 Finally, the use of conditioning regimens that include cyclophosphamide, ATGAM and cyclosporine with or without methotrexate have decreased the incidence of graft failure to 0-5%.^31,32,39

Bacigalupo et al³⁴ also looked at factors that are likely to be predictive of better outcome in patients treated with IST. Three subgroups of patients were analyzed according to neutrophil count: 0-0.2 ´ 10⁹/l, 0.21-0.5 ´ 10⁹/l, and greater than 0.5 ´ 10⁹/l. The actuarial 5-year survival rates for patients treated after 1990 and stratified in these three risk groups were 70%, 78% and 77%, respectively. Furthermore, they reported that a combination of young age and low neutrophil count favor BMT. For example, a 10 year old with a neutrophil count of 0.2 ´ 10⁹/l would, according to them, have a 14% 5-year failure-free survival difference in favour of BMT. Our data are in keeping with these findings as ANC at diagnosis was the only factor in our study that was predictive of outcome in the IST patients (P = 0.0058). One year after diagnosis, the estimated survival of those patients with ANC <0.2 ´ 10⁹/l at diagnosis is 25 ± 15% compared with 93 ± 6% for those with ANC >0.2 ´ 10⁹/l.

One of the negative aspects of successful bone marrow transplantation, however, is the occurrence of second malignancies. The malignant tumors most often encountered include solid tumors, epidermoid carcinomas, and leukemias. The combined Seattle and Paris long-term follow-up study reported second malignancy incidence rates of 14% at 10 years.⁴⁰ This incidence was higher than the overall rate of 3.1% reported by the EBMT-SAA.⁴¹ In both these studies, significant risk factors for the development of a second malignancy included male sex, age at diagnosis and use of irradiation in the conditioning regimen. Among our patients, the only documented second malignancy post-transplantation, was in a 13-year-old male who had not received any radiation as part of his conditioning regimen and developed AML 34 months after transplant.

Immunosuppressive therapy in children with severe aplastic anemia has traditionally been reserved for those without a suitable sibling donor. Survivals for patients have ranged from 30 to 70%^42,43 and the results have improved significantly in the past decade as seen in our previously published experience in the 1970s and 1980s. These improvements are due to the advent of ATGAM, cyclosporine and regimens that optimize combination immunosuppressive therapy as well as the improvements in transfusion support, antimicrobial therapy and general supportive care. In a randomized controlled trial of antithymocyte globulin vs cyclosporine, Gluckman et al⁴⁴ showed overall survival of 66% at 2 years for both therapies. The German prospective randomized trial reported on a comparison of ATG and methylpredniosolone vs ATG, methylprednisolone and cyclosporine with significantly higher response rates at 3 months for the group receiving cyclosporine (65% vs 34%) and similar survival rates (61% vs 54%).⁴⁵ The modification of immunosuppressive therapy has led to significant improvement in survival with overall 5-year actuarial survival in all categories of patients reported as 43% for those treated before 1984 and 66% for those treated after 1984 (P = 0.0001). Our study confirms this finding with 65% survival (median follow-up of 30 months for all 20 patients; 36 months for the 14 IST patients alive at last follow-up) for those treated between 1987 and 1997 vs 33% (median follow-up of 24 months) for those treated between 1977 and 1987.

Graft failure, chronic GVHD, chronic organ dysfunction due to conditioning regimens and secondary malignancies can complicate marrow transplantation, but recurrent aplasia and the evolution of clonal diseases can complicate immunosuppressive therapy. In the EBMT-SAA registry,⁴⁶ 74 of 358 patients (35.2%) relapsed at a mean of 778 days after initial therapy. The 10-year cumulative rate of cancer was 18.8% after immunosuppression and 3.1% after transplantation. Of 42 malignant conditions after immunosuppressive therapy, 19 were myelodysplastic, 15 acute leukemia, one non-Hodgkin's lymphoma, and seven solid tumors.

Although bone marrow transplantation has been considered first-line treatment in young patients with severe aplastic anemia who had an HLA-identical sibling donor, this approach has been called into question by recent retrospective studies suggesting that intensive immunosuppressive therapy may produce equivalent rates of transfusion independence and survival.^12,13 There are few direct comparisons of these two approaches. The published studies, like ours, are retrospective with data collected from single institutions or large data registries. Furthermore, most studies have small numbers and usually span a long period of time during which significant changes occurred in immunosuppressive regimens, transplantation protocols, and supportive care, making the results difficult to interpret.

For example, during the late 1970s and 1980s, cyclosporine was not a routine part of immunosuppressive regimens; hence the results may not reflect those currently obtained with intensive immunosuppressive protocols. Bayever et al⁹ compared the 2-year survival of 35 patients less than 25 years of age treated with transplantation to 22 patients treated with antithymocyte globulin therapy and reported a survival of 72% for the transplant group vs 45% for ATG therapy (P = 0.18). Between 1970 and 1986, the EBMT results comparing transplantation to immunosuppressive therapy with various antithymocyte globulin protocols without cyclosporine reported survival of 66% in the transplant group vs 56% in the other group.⁴⁶ Doney et al⁴⁸ compared outcomes of transplantation to antithymocyte globulin, corticosteroids and androgens between 1978 and 1991 and reported that marrow transplantation was the preferred treatment for younger patients with acquired aplastic anemia with matched related donors. In patients less than 6 years of age, the survival rate at 15 years was 100% for those treated with bone marrow transplantation and 51% for the 25 patients treated with immunosuppressive therapy (P = 0.0006). For patients between 6 and 19 years of age, survival was again significantly better with BMT (P = 0.001).

Lawlor et al¹² reported equivalent survival rates and rates of transfusion independence in 27 patients consecutively treated in a single institution between 1982 and 1994. At a median follow-up of 33.5 months for the immunosuppression group (n = 18) and 54 months for the transplant group (n = 9), actuarial survival was 75% and 92% for transplantation and immunosuppression groups, respectively (P = 0.15). The immunosuppression regimen in this small group consisted of ATG and steroids. Fifteen patients also received cyclosporine. Gillio et al¹³ reviewed 48 patients consecutively treated between 1983 and 1992 at the Memorial Sloan-Kettering Cancer Center and reported equivalent survival rates of 75.6% and 73.8%, respectively, at 120 months. It was postulated that the improvement in survival of the immunosuppression group was secondary to the addition of cyclosporine and the improvement in supportive care.

In this study, we compared survival of patients with severe aplastic anemia treated with transplantation vs immunosuppression over the past decade but also took into consideration the modifications in treatment for both modalities by comparing outcomes with our published data from the same institution from a decade earlier.¹¹ We clearly showed an improvement in outcome in both modalities in the recent 10-year timeframe.

However, our findings differ from those of Lawlor et al¹² and Gillio et al¹³ regarding the equivalence in survival and transfusion independence of both treatment approaches. In the 20 patients treated with immunosuppression between 1987 and 1997, 14 (70%) are transfusion independent and, at 3 years, the estimated survival for the IST group is 65 ± 13% (median follow-up 3 years; range: 0.2-6.7 years) compared with 95 ± 5% for the BMT-MRD group (median follow-up 5.9 years; range: 1.4-9 years) and 100% for the BMT-MUD group (median follow-up 4.3 years; range: 3.3-6.7 years). Unfortunately, data regarding quality of life issues were not consistently available in our cohort to make a meaningful comparison possible.

In conclusion, our results demonstrate that survival of pediatric patients with severe aplastic anemia has improved significantly over the past decade regardless of whether they receive transplantation or immunosuppression. However, unlike the recent reports of equivalent survival by Lawlor et al,¹² and Gillio et al,¹³ our data demonstrate a significant survival advantage for bone marrow transplantation. Because of the inherent limitations of most of these retrospective studies, inadequate long-term follow-up data to analyze rates of relapse, second malignancy and clonal disease, and the excellent recent marrow transplantation results in our study as well as others, we would recommend the use of bone marrow transplantation as first-line therapy for children with severe aplastic anemia who have an HLA-matched sibling donor. However, it is clear that a randomized controlled trial comparing the two approaches is needed to provide a definitive answer regarding the relative merits of these two treatment modalities, as well as to document the long-term quality of life issues, which may affect the choice of therapy.

1 The International Agranulocytosis and Aplastic Anemia Study. Incidence of aplastic anemia: the relevance of diagnostic criteria. Blood 1987; 70: 1718-1721, MEDLINE

2 Young NS, Alter BP. Definitive treatment of acquired aplastic anemia. In: Young NS, Alter BP (eds). Aplastic Anemia: Acquired and Inherited. WB Saunders Co: Philadelphia, 1993 pp 159-200,

3 Thomas ED, Storb R, Fefer A et al. Aplastic anemia treated by bone marrow transplantation. Lancet 1972; 1: 284-289, MEDLINE

4 Camitta BM, Thomas ED, Nathan DG et al. A prospective study of androgens and bone marrow transplantation for treatment of severe aplastic anemia. Blood 1979; 53: 504-514, MEDLINE

5 Mathé G, Amiel JL, Schwarzenberg L et al. Bone marrow graft in man after conditioning by antilymphocytic serum. Br Med J 1970; 2: 131-136, MEDLINE

6 Speck B, Gluckman E, Haak HL et al. Treatment of aplastic anemia by antithymocyte globulin with and without allogeneic bone marrow infusions. Lancet 1977; 2: 1145-1148, MEDLINE

7 Speck B, Gratwohl A, Nissen C et al. Treatment of severe aplastic anemia. Exp Hematol 1986; 14: 126-132, MEDLINE

8 Young N, Griffith P, Brittain E et al. A multicenter trial of antithymocyte globulin in aplastic anemia and related diseases. Blood 1988; 72: 1861-1868, MEDLINE

9 Bayever E, Champlin R, Ho W et al. Comparison between bone marrow transplantation and antithymocyte globulin in treatment of young patients with aplastic anemia. J Pediatr 1984; 105: 920-925, MEDLINE

10 Locasciulli A, van't Veer L, Bacigalupo A et al. Treatment with marrow transplantation or immunosuppression of childhood acquired severe aplastic anemia: a report from the EBMT SAA Working Party. Bone Marrow Transplant 1990; 6: 211-217, MEDLINE

11 Halperin DS, Grisaru D, Freedman MH et al. Severe aplastic anemia in children: 11-year experience with bone marrow transplantation and immunosuppression therapy. Am J Pediatr Hematol Oncol 1989; 11: 304-309, MEDLINE

12 Lawlor ER, Anderson RA, Davis DH et al. Immunosuppressive therapy: a potential alternative to bone marrow transplantation as initial therapy for acquired severe aplastic anemia in childhood? J Pediatr Hematol Oncol 1997; 19: 115-123, MEDLINE

13 Gillio A, Boulad F, Small T et al. Comparison of long-term outcome of children with severe aplastic anemia treated with immunosuppression versus bone marrow transplantation. Biol Blood Marrow Transplant 1997; 3: 18-24, MEDLINE

14 Bacigalupo A, Broccia G, Corda W et al. Antilymphocyte globulin, cyclosporin, and granulocyte colony-stimulating factor in patients with acquired severe aplastic anemia (SAA): a pilot study of the A+EBMT SAA working party. Blood 1995; 5: 1348-1353,

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

16 Peto R, Pike MC, Armitage P et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient, II: analysis and examples. Br J Cancer 1977; 35: 1-39, MEDLINE

17 Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22: 719-748,

18 Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. Wiley: New York, 1980, p 169.

19 Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat 1988; 16: 1141-1154,

20 Sanders JE, Whitehead J, Storb R et al. Bone marrow transplantation experience for children with aplastic anemia. Pediatr 1986; 77: 179-186,

21 Gluckman E, Socie G, Devergie A et al. Bone marrow transplantation in 107 patients with severe aplastic anemia using cyclophosphamide and thoraco-abdominal irradiation for conditioning: long-term followup. Blood 1991; 78: 2451-2455, MEDLINE

22 Bortin MM, Gale RP, Rimm AA. Allogeneic bone marrow transplantation for 144 patients with severe aplastic anemia. JAMA 1981; 245: 1132-1139, MEDLINE

23 Hows JM, Palmer S, Gordon-Smith EC. Use of cyclosporin A in allogeneic bone marrow transplantation for severe aplastic anemia. Transplantation 1982; 33: 382-386, MEDLINE

24 Storb R, Doney KC, Thomas ED et al. Marrow transplantation with or without donor buffy coat cells for 65 transfused aplastic anemia patients. Blood 1982; 59: 236-246, MEDLINE

25 Feig SA, Champlin R, Arensen E et al. Improved survival following bone marrow transplantation for aplastic anemia. Br J Haematol 1983; 54: 509-517, MEDLINE

26 Ramsay WKC, Kim TH, Nesbit ME et al. Total lymphoid irradiation and cyclophosphamide as preparation of bone marrow transplantation in severe aplastic anemia. Blood 1980; 55: 344-346, MEDLINE

27 Storb R, Deeg HJ, Farewell V et al. Marrow transplantation for severe aplastic anemia: methotrexate alone compared with a combination of methotrexate and cyclosporin for prevention of acute graft versus host disease. Blood 1986; 69: 119-126,

28 Gluckman E, Devergie A, Meletis J et al. Bone marrow transplantation for severe aplastic anemia: report of 97 consecutive patients. Bone Marrow Transplant 1987; 2: 101-104,

29 Anasetti C, Storb R, Longton G et al. Donor buffy coat cell infusion after marrow transplantation for aplastic anemia. Blood 1988; 72: 1099-1100, MEDLINE

30 Storb R, Deeg HJ, Pepe M et al. Graft-versus-host disease prevention by methotrexate combined with cyclosporin compared to methotrexate alone in patients given marrow grafts for severe aplastic anemia: long-term followup of a controlled trial. Br J Haematol 1989; 72: 567-572, MEDLINE

31 Bunin N, Leahey A, Kamani N et al. Bone marrow transplantation in pediatric patients with severe aplastic anemia: cyclophosphamide and antithymocyte globulin conditioning followed by recombinant human granulocyte-macrophage colony stimulating factor. J Pediatr Hematol Oncol 1996; 18: 68-71, MEDLINE

32 Storb R, Etzioni R, Anasetti C et al. Cyclophosphamide combined with antithymocyte globulin in preparation for allogeneic bone marrow transplants in patients with aplastic anemia. Blood 1994; 84: 941-949, MEDLINE

33 May WS, Sensenbrenner LL, Burns WH et al. BMT for severe aplastic anemia using cyclosporin. Bone Marrow Transplant 1993; 11: 459-464, MEDLINE

34 Bacipalupo A, Brand R, Oneto R et al. Treatment of acquired severe aplastic anemia: bone marrow transplantation compared with immunosuppressive therapy - the European group for blood and marrow transplantation experience. Semin Hematol 2000; 37: 69-80, MEDLINE

35 McCann SR, Bacigalupo A, Gluckman E et al. Graft rejection and second bone marrow transplants for acquired aplastic anemia: a report form the aplastic anemia working party of the European Bone Marrow Transplant Group. Bone Marrow Transplant 1994; 13: 233-237, MEDLINE

36 Gluckman E, Horowitz MM, Champlin RE et al. Bone marrow transplantation for severe aplastic anemia: influence of conditioning and graft-versus-host disease prophylaxis regimens on outcome. Blood 1992; 79: 269-275, MEDLINE

37 Champlin RE, Horowitz MM, van Bekkum DW et al. Graft failure following bone marrow transplantation for severeaplastic anemia: risk factors and treatment results. Blood 1989; 73: 606-613, MEDLINE

38 Sanders JE, Storb R, Anasetti C et al. Marrow transplant experience for children with severe aplastic anemia. Am J Pediatr Hematol Oncol 1994; 16: 43-49, MEDLINE

39 Horstmann M, Stockschlader M, Kabisch H et al. Cyclo-phosphamide/antithymocyte globulin conditioning of patients with severe aplastic anemia transplanted with bone marrow from HLA-identical related donors. Ann Intern Med 1997; 126: 116-122, MEDLINE

40 Deeg HJ, Socie G, Shoch G et al. Malignancies after marrow transplantation for aplastic anemia and Fanconi anemia: a joint Seattle and Paris analysis of results in 700 patients. Blood 1996; 87: 386-392, MEDLINE

41 Socie G, Henry-Amar M, Bacigalupo A et al. Malignant tumours occurring after treatment of aplastic anemia. New Engl J Med 1993; 329: 1152-1157, MEDLINE

42 Hinterberger-Fischer M, Hocker P, Lechner K et al. Oral cyclosporin A is effective treatment for untreated and also for previously immunosuppressed patients with severe bone marrow failure. Eur J Haematol 1989; 43: 136-142, MEDLINE

43 Leonard EM, Raefsky E, Griffith P et al. Cyclosporin therapy of aplastic anemia, congenital and acquired red cell aplasia. Br J Haematol 1989; 72: 278-284, MEDLINE

44 Gluckman E, Esperou-Boudreau H, Baruchel A et al. Multicenter randomized study comparing cyclosporin A alone and antithymocyte globulin with prednisone for the treatment of severe aplastic anemia. Blood 1992; 79: 2540-2546, MEDLINE

45 Frickhofen N, Kaltwasser JP, Schrezenmeier H et al. Treatment of aplastic anemia with antilymphocyte globulin and methylprednisolone with or without cyclosporine. New Engl J Med 1991; 324: 1297-1304, MEDLINE

46 Schrezenmeier H, Marin P, Raghavachar A et al. Relapse of aplastic anemia after immunosuppressive treatment: a report from the European Bone Marrow Transplantation Group SAA Working Party. Br J Haematol 1993; 85: 371-377, MEDLINE

47 Bacigalupo A, Hows J, Gluckman E et al. Bone marrow transplantation versus immunosuppression for the treatment of severe aplastic anaemia (SAA): a report of the EBMT SAA Working Party. Br J Haematol 1988; 70: 177-182, MEDLINE

48 Doney K, Leisenring W, Storb R et al. Primary treatment of acquired aplastic anemia: outcomes with bone marrow transplantation and immunosuppressive therapy. Ann Intern Med 1997; 126: 107-115, MEDLINE

Figure 1 Cumulative incidence distributions of time to CR by treatment group.

Figure 2 Kaplan-Meier plot of overall survival by treatment group.

Table 1  Patient characteristics at the time of diagnosis of SAA

Table 2  Distribution of response by treatment group

Table 3  Main complications

Table 4  Causes of death