Bone Marrow Transplantation, (1999) 24, 1299–1306  1999 Stockton Press All rights reserved 0268–3369/99 $15.00 http://www.stockton-press.co.uk/bmt Second marrow transplants for graft failure in patients with thalassemia

D Gaziev, P Polchi, G Lucarelli, M Galimberti, P Sodani, E Angelucci, C Giardini, D Baronciani and B Erer

Divisione di Ematologia e Centro Trapianti di Midollo Osseo di Muraglia, Azienda Ospedaliera di Pesaro, Pesaro, Italy

Summary:

Thirty-two thalassemic patients with a median age of Currently, the only curative treatment for thalassemia is 7.7 years (range 3.4–26 years) were given a second bone marrow transplantation.1–3 Graft failure owing to HLA-identical related marrow transplant (BMT2) for rejection and other causes is one of the more common graft failure. Four patients were in class 1 and 28 obstacles to successful marrow transplantation, especially patients in classes 2 and 3. Twenty-one patients had full in the non-malignant BMT setting. Unlike the situation with thalassemia recurrence (first group) and 11 patients had hematological malignancies, patients with non-malignant aplastic marrows (second group) either with or without diseases such as aplastic anemia and thalassemia have an residual donor marrow cells after the first BMT increased risk of graft failure after BMT. The incidence of (BMT1). As conditioning regimen for BMT2 all but five rejection after BMT for thalassemia varies from 7% (in risk patients received BUCY or CY in association with total class 1 and class 2 patients) to 22% (in class 3 patients),4,5 lymphoid irradiation (TLI) and/or anti-lymphocyte while its incidence in adult patients is low, 4%.6 In class globulin (ALG), whereas nine patients received a new 3 patients analysis of rejection has revealed that patients preparative regimen with hydroxyurea, azathioprine, who received less than 100 transfusions, regimens contain- fludarabine before conditioning with BUCY. Twenty ing less than 200 mg/kg CY and who showed elevated one of 31 evaluable patients (67.7%) had initial, and 16 aspartate aminotransferase had an increased rate of rejec- (51.6%) had sustained engraftment. Ten patients tion.5 In patients given HLA-identical related marrow trans- (32.3%) failed to engraft. Overall and event-free sur- plants for thalassemia, graft failure is usually accompanied vival for the entire group of patients were 49% and by autologous recovery of the thalassemic marrow.3–5 How- 33%, respectively, with a median follow-up of 4 years ever, in a small group of patients graft failure may not (range 0.6–14 years) for surviving patients. Event-free be accompanied by thalassemia recurrence after marrow survival was higher in the second group of patients com- transplantation, which is a life-threatening situation. pared with the first group (41% vs 29%). The second Two types of graft failure, primary (without evidence of group of patients appeared to have less graft failure engraftment or hematologic recovery of donor cells) and compared with the first group (30% vs 63%; P = 0.1). late (losing the graft with autologous recovery or re-occur- Transplant-related mortality was 28%. A linear step- rence of aplastic marrow), have been recognized in patients wise regression analysis revealed that occurrence of with aplastic anemia.7 Graft failure may be immunolog- graft failure within 60 days after BMT1 (P = 0.04) and ically mediated (graft rejection), or related to a low stem absence of residual donor marrow cells (P = 0.009) pre- cell inoculum, cytomegalovirus infection, GVHD, drug tox- dicted for graft failure following BMT2, whereas the icity, leukemic relapse or other causes.8 Treatment for graft occurrence of graft failure after 60 days (P = 0.03) had rejection or failure may be supportive care or second mar- a positive influence on survival following BMT2. The row transplant. Therapy with cytokines has improved the incidence of grade у2 acute GVHD was low (14%). outcome of patients with graft failure accompanied by neu- Eight of nine patients who received the new preparative tropenia. However, results of either cytokine therapy or regimen are alive, four without thalassemia. This study second transplant are poor.8,9 For patients with primary shows that BMT2 can be an effective therapy for a pro- graft failure accompanied by an aplastic marrow who are portion of patients with poor survival expectancies unresponsive to cytokine therapy, second marrow trans- despite conventional treatment. plantation is the only salvage approach. Keywords: second BMT; graft failure; thalassemia; Although conventional treatment has dramatically regimen related toxicity; conditioning improved survival and quality of life in patients with thalas- semia, this remains a progressive disease with therapy- related complications progressing with time. Bone marrow transplantation today is the only curative therapy that allows patients with thalassemia to lead a normal life. Thus, Correspondence: Dr D Gaziev, Divisione di Ematologia e Centro Trapianti di Midollo Osseo di Muraglia, Azienda Ospedaliera di Pesaro, Via patients with thalassemia who experience graft failure after Lombroso, 61100 Pesaro, Italy BMT should be offered a second transplant to cure their Received 13 May 1999; accepted 22 July 1999 disease. Despite a high rejection rate, especially in inad- Second BMT for graft failure in thalassemia D Gaziev et al 1300 equately iron-chelated thalassemic patients with liver dam- bone marrow from the same donor. For one of these four age, there are few reports of second transplants for thalasse- patients the donor for BMT1 was the haploidentical mother mia.10,11 In a preliminary study we analyzed the results of and for BMT2 it was a HLA-identical sibling. Patients were second marrow transplantation for thalassemia.12 In the categorized into two groups according to the outcome of present retrospective study we have analyzed the outcome graft failure after BMT1. The first group included 21 of second BMT in 32 patients who developed graft failure patients who had complete recovery of thalassemic marrow after a first transplant for thalassemia. The objectives of and were receiving conventional treatment before the this study were: (1) to examine survival and graft failure second transplant. Eleven patients who developed aplasia in patients after a second BMT; and (2) to identify factors after BMT1 were allocated to the second group. that predispose patients to graft failure after a second BMT. Definition of graft failure Patients and methods Graft failure was defined as primary or secondary failure. Primary graft failure is characterized by the absence of allo- Ͼ Patients geneic recovery by day 21 with or without autologous reconstitution. Secondary graft failure was defined as early From October 1982 to November 1998, 32 patients with and late failure (losing the graft after transient engraftment thalassemia received a second HLA-identical related mar- within 60 days and after 60 days respectively). Secondary row transplant (BMT2) for graft failure. One of these graft failure may be accompanied by thalassemia recurrence received a third transplant for primary graft failure after or marrow aplasia. BMT2 from her haploidentical mother. Another patient received a granulocyte colony-stimulating factor (G-CSF) Conditioning regimens and GVHD prophylaxis mobilized peripheral blood stem cell boost for graft failure following BMT2 without signs of re-engraftment. Seven As conditioning regimen for BMT2 among 21 patients in patients received their first transplant (BMT1) elsewhere. the first group, three patients received BUCY only, seven Characteristics of these 32 patients are shown in Table 1. BUCY and total lymphoid irradiation (TLI), two BUCY There were 14 males and 18 females with a median age of plus anti-lymphocyte globulin (ALG, Merrieux, France), 7.7 years (range 3.4–26 years). Four patients were in class and the remaining nine patients received a new preparative 1, 14 in class 2 and 14 in class 3. Marrow donors and regimen including a cytoreductive-immunosuppressive and recipients were HLA-A, HLA-B, HLA-C and DR identical hypertransfusion-intensive chelation therapy before BUCY siblings in 27 patients, parents in four patients and a cousin to reduce the hyperplastic bone marrow and increase in one patient. Mixed lymphocyte cultures were mutually immunosuppression before BMT2 (Table 2). From day −45 nonreactive in all patients. All but four patients received to day −12 these nine patients received hydroxyurea 30 mg/kg/day, azathioprine 3 mg/kg/day along with hyper- transfusions and i.v. continuous desferrioxamine 40 − − Table 1 Characteristics of the patients before BMT1 and BMT2 mg/kg/day and from day 17 to day 13 fludarabine (Fludara; Schering AG, Berlin, Germany) 20 mg/m2/day as BMT1 BMT2 outpatients. They were given BUCY beginning on day −9. Four of the second group of patients were given CY plus Median patient age, years (range) 5.2 7.7 TLI and ALG, three BUCY in association with TLI and/or (1.6–25) (3.4–26) ALG, one CYTLI, one CYALG and two patients ALG Median donor age, years (range) 12 13 only. Total lymphoid irradiation was given on day −1. The (4–43) (1.6–43) 13 Patient sex, M/F 14/18 14/18 dose of radiation was 750 cGy administered at 25 Donor sex, M/F 17/15 19/13 cGy/min in a single dose through anterior and posterior Patient–donor sex match 10 8 fields using a 10 MV linear accelerator. Lungs, thyroid, Risk class: class 1/class 2/class 3 4/14/14 liver, kidneys and testicles in males were shielded. As Ͼ Hepatomegaly 2 cm, n 19 26 GVHD prophylaxis class 1 and 2 patients received CsA + Splenomegaly Ͼ2 cm, n 810 + Splenectomy 9 11 methylprednisolone (MP) and class 3 patients CsA short 4,5 Chronic active hepatitis 9 5 MTX + MP as previously described. Acute and chronic Chronic persistent hepatitis 5 9 GVHD were diagnosed according to the Seattle criteria.14 Liver fibrosis, grade 1–3 14 18 Patients received the marrow infusion 36 h after the last HBAg: negative/positive 31/1 31/1 dose of CY, on day 0. The median marrow cell dose infused HCV-Ab: negative/positive 19/6 19/5 8 8 Median bilirubin, mg/dl (range) 0.75 0.7 was 3.9 × 10 /kg (range 1.9–8.0 × 10 /kg). The median (0.3–2.4) (0.3–2.0) interval from BMT1 to graft failure was 60 days (range Median ALT, U/l (range) 42 46 15–610 days), and the median interval from BMT1 to (8–545) (9–995) BMT2 was 675 days (35–3794 days) (Table 2). Patients or Median AST, U/l (range) 30 32 (13–365) (8–372) their legal guardians signed informed consent. Median number of transfused RBC, 70 87 units (range) (8–613) (18–630) Evaluation of graft status Median serum ferritin, ng/ml (range) 1478 2357 (411–7080) (694–18696) Globin-chain synthesis of marrow and peripheral blood reti- culocytes were examined by incorporation of 3H-leucine Second BMT for graft failure in thalassemia D Gaziev et al 1301 Table 2 Characteristics of BMT and engraftment

BMT1 BMT2

Patients n Values Patients n Values

Median time (days) to graft failure after BMT 32 60 (15–610) 15 27 (20–90) Median time (days) of interval between BMT1 and 32 675 (35–3794) BMT2 Conditioning 20 BU14CY200 3 BU14CY200 7 BU16CY120/160 1 BU8CY120ALG 1 BU8CY200TBI300 2 BU14CY200/160ALG 3 BU14CY120/160ALG 1 BU10CY120ALGTLI 1 BU16CY200ALG 2 BU16CY120TLI 5 BU14CY160/200TLI 1 BU16CY120TLI 2a BU14CY90a 7a BU14CY200a 2 ALG 1 CY200ALG 1 CY200TLI 4 CY200ALGTLI GVHD prophylaxis 3 MTX 1 MTX 4 CsA 1 CsA 10 CsA + MP 14 CsA + MP 15 CsA + MTX + MP 16 CsA + MTX + MP Median marrow cell dose infused × 108/kg 2.8 (1.5–10) 3.9 (1.9–8.0) Median time (days) to granulocyte count of 0.5 × 109/l 27 20 (14–40) 21 19 (11–63) Median time (days) to granulocyte count of 1.0 × 109/l 26 23 (16–80) 21 23 (22–63) Median time (days) to platelet count of 20 × 109/l 26 18 (8–46) 21 17 (3–83) Median time (days) to platelet count of 50 × 109/l 23 24 (10–100) 16 25 (12–90) aThese patients received hydroxyurea, azathioprine and fludarabine along with hypertransfusions and i.v. continuous desferrioxamine before conditioning with BUCY. BU = busulfan; CY = cyclophosphamide; ALG = anti-lymphocyte globulin; TLI = total lymphoid irradiation; CsA = cyclosporine; MP = methylpredniso- lone; MTX = methotrexate. followed by column or high-pressure liquid chromato- method of Kaplan and Meier.15 Associations between sur- graphy (HPLC). Chimerism status was studied by in situ vival, graft failure and the pre-transplant variables (age, sex Y chromosome hybridization of bone marrow or blood mismatch, pre-transplant transfusions, serum ferritin, liver samples in sex-mismatched donor–recipient pairs and by function tests, chronic active hepatitis, liver fibrosis, risk analysis of variable number tandem repeat (VNTR) poly- classes, spleen size, interval from first BMT to graft failure, morphism in bone marrow and/or blood samples in sex- interval between transplants, mixed chimerism, HLA serol- matched pairs. ogy, marrow cell dose, dose of cyclophosphamide, TLI, ALG) were tested in univariate analyses using the Pearson chi-squared statistics. Variables significant at the P Ͻ 0.1 Supportive care level were assessed in stepwise linear regression analyses Patients were maintained in strict isolation in single rooms (backward elimination) using the Statistix statistical pack- with positive-pressure HEPA-filtered air and received oral age.16 This package was also used for engraftment charac- broad spectrum nonabsorbable antibiotics. Prophylactic teristics. The results were analyzed on 31 March 1999. broad spectrum systemic antibiotics were begun on day −1 and continued until the granulocyte level exceeded 0.5 × 109/l. From 1987, all patients received i.v. ampho- tericin B (0.3 mg/kg/day) as antifungal prophylaxis. Results Patients were given acyclovir (15 mg/kg/day) as herpes viruses prophylaxis (days −1to+60) and trimethoprim/ sulfamethoxazole (5 mg/kg/day trimethoprim) before (days Patients −9to−1) and after BMT (2 days per week) for prophylaxis of Pneumocystis carinii. All blood products were irradiated Four patients (12.5%) had primary and 28 patients (87.5%) with 30 Gy. had secondary graft failure following BMT1. In the first group of patients three had primary, nine had secondary Statistical analysis early and nine had secondary late graft failure. Among the second group of patients, one had primary, two had second- The endpoints were overall survival, event-free survival ary early and eight had secondary late graft failure. Detec- and graft failure or rejection. The probability of survival, tion of cytomegalovirus (CMV) was performed in eight of event-free survival and rejection were estimated using the the second group patients with negative results in all cases. Second BMT for graft failure in thalassemia D Gaziev et al 1302 Engraftment 1 Thirty-one of 32 patients were evaluable for engraftment 0.8 Survival following BMT2. One patient died (at day 13) before 0.6 engraftment was evaluable. Twenty-one of 31 patients 55% Event-free survival (67.7%) showed allogeneic engraftment and five of them 0.4

Probability 29% subsequently lost their grafts with autologous reconsti- 0.2 tution. Thus, 16 patients (51.6%) had sustained engraftment. Ten of 31 patients (32.3%) had failed to 0 engraft. Three of four patients with primary and 13 of 28 024681012 patients with secondary graft failure occurring after BMT1 Years had sustained engraftment following BMT2. Two of four Figure 2 Overall and event-free survival for 21 patients who had com- patients given BMT2 from different donors had sustained plete thalassemia recurrence after BMT1 (first group) and received a engraftment. Nine of the first group (42.8%) and seven of second BMT. 10 evaluable for engraftment patients of the second group = (70%) had sustained engraftment after BMT2 (P 0.09). probability of event-free survival was higher (41%) in the Ͼ × 9 The median times to granulocyte count of 0.5 10 /l and second group of patients as compared to the first group Ͼ × 9 platelet count of 20 10 /l were comparable after BMT1 (29%) (Figures 2 and 3). The probability of graft failure and BMT2 (Table 2). was 30% in the second group and 63% in the first group of patients (P = 0.1). Eleven of 17 patients (64.7%) who Pattern of graft failure after BMT2 developed graft failure within 60 days, and four of 14 patients (33%) developing graft failure over 60 days after In the first group seven patients developed primary, three BMT1 had graft failure following BMT2 (P = 0.04). Over- patients secondary early and two patients secondary late all and event-free survival after BMT2 were better in failure. Three patients in the second group had primary patients who had graft failure over 60 days compared with graft failure. Three of 10 patients with primary graft failure patients who developed graft failure within 60 days follow- had persistent aplasia, whereas seven patients had full thal- ing BMT1 (85.7% and 52.9% vs 57.1% and 17.6%, respect- assemia recurrence. All five patients with secondary graft ively; P = 0.05 for overall survival and P = 0.02 for event- failure had complete regrowth of thalassemic marrow. free survival). Two of four patients who developed primary and nine of 28 patients who had secondary graft failure GVHD after BMT1 are surviving without thalassemia following BMT2. One of 27 evaluable patients after BMT1 developed grade Univariate analyses showed that the occurrence of graft 2 acute GVHD which resolved completely before graft fail- failure within 60 days (vs Ͼ60 days) after BMT1 (P = ure had occurred. Chimerism study in this patient showed 0.044) and the absence of residual donor marrow cells (vs no residual donor cells. After BMT2 three of 21 evaluable = у the presence) (P 0.010) were associated with a high inci- patients (14%) developed grade 2 acute GVHD. Only one dence of graft failure following BMT2. These results were of 11 evaluable patients developed moderate chronic confirmed by a stepwise regression analysis with the same GVHD. levels of significance. Patients who received less than 200 mg CY as conditioning therapy appeared to have more fail- Survival and graft failure ure compared with the patients receiving 200 mg CY (75% vs 38%; P = 0.07). Univariate and stepwise regression Overall and event-free survival for the entire group of analysis for survival showed that the occurrence of graft patients are shown in Figure 1. Event-free survival was failure over 60 days (vs Ͻ60 days) after BMT1 (P = 0.033) 37% at 2 years and 33% at 13 years with a median follow- had a positive influence on survival following BMT2. Fac- up for surviving patients of 4 years (range 0.6 to 14 years). tors not influential on rejection and survival after BMT2 The probability of graft failure in all patients was 42%. The were: age, sex mismatch, pre-transplant transfusions, serum

1 1

0.8 0.8

0.6 Survival 0.6 49% Survival 0.4 Event-free survival 0.4 41% 33% Event-free survival Probability Probability 0.2 0.2

0 0 02468101214 02468101214 Years Years

Figure 1 Overall and event-free survival for 32 patients receiving a Figure 3 Overall and event-free survival for 11 patients who developed second BMT for thalassemia. aplasia after BMT1 (second group) and received a second BMT. Second BMT for graft failure in thalassemia D Gaziev et al 1303 ferritin, liver function tests, chronic active hepatitis, liver Chimerism and outcome of BMT2 fibrosis, risk classes, spleen size, interval between BMT1 and BMT2, HLA serology, dose of cyclophosphamide and Twenty-seven patients had data regarding the origin of the use of TLI or ALG in conditioning regimen. residual donor cells prior to BMT2. In six patients it was determined by conventional cytogenetic examination, and in the remaining patients by VNTR examination of DNA Results of BMT2 after a new conditioning regimen and/or FISH. Twenty patients had no residual donor cells, Nine patients received a new preparative regimen with pre- while seven patients showed persistence of cells of donor conditioning cytoreductive-immunosuppressive and hyper- origin in bone marrow before BMT2. In the first group only transfusion-intensive chelating therapies. Results of BMT2 one of 18 patients evaluable for chimerism (5.5%) had after this regimen are shown in Table 3. Seven of nine residual donor cells, while six of nine evaluable patients of patients had engraftment and three of them subsequently the second group (66.6%) had residual donor cells ranging lost their grafts. Two patients developed primary graft fail- from 1% to 100% (on FISH or VNTR-DNA examinations) ure and one (UPN 1287) with an aplastic marrow was given before BMT2. The only patient who had chiefly donor cells a third BMT from her haploidentical mother after con- (100%) prior to BMT2 had a severe aplastic marrow not ditioning with CY200 plus ALG. This patient had prompt responsive to CsA and steroids for more than 2 months. engraftment but died of severe acute GVHD. All seven All seven patients with residual donor cells had secondary engrafted patients showed fast hematological recovery. The late graft failure and sustained engraftment, whereas nine median times to granulocyte count (n = 7 patients) of of 20 patients (45%) without donor cells had sustained Ͼ0.5 × 109/l and to platelet count of Ͼ20 × 109/l were 15 engraftment following BMT2. Three of seven patients who days (range 12–21 days) and 9 days (range 9–22 days), had residual donor marrow cells are alive without thalasse- respectively. Only one patient developed moderate mia, whereas 15 of 20 patients without residual cells are hemorrhagic cystitis and two patients had cytomegalovirus alive and eight of them are free of thalassemia. reactivation without clinical disease.

Table 3 Characteristics of patients who received a new preparative regimen for BMT2

UPN 1157 1052 928 391 1180 1287 961 1296 1023

BMT1 Age, years 2 2 14 10.4 2.5 5.2 3 2.5 10 Sex MFFFFFFMF Liver size, cm 1 1 5 5 1 0 1 2 2 Spleen size, cm 1 0 out 2 0 0 0 2 2 Pre-transplant transfusions, units 30 20 250 120 25 15 50 15 140 Serum ferritin, ng/ml 1285 574 5331 6738 787 1000 1037 411 714 Chronic active hepatitis — no yes no — — no no no Liver fibrosis, grade — 0 3 2 — — 0 0 1 Conditioning regimen BuCy BuCy BuCy BuCy BuCy BuCy BuCy BuCy BuCy GVHD prophylaxis CSMP CSMP CSMP CSMP CSMP CSMP CSMP CSMP CSMP MTX MTX Acute GVHD, grade 2 1 0 0 0 0 0 0 0 Days of thalassemia recurrence 36 417 396 51 606 15 229 48 61

BMT2 Age 3.4 5 18 20.8 4.4 7.5 8.2 5.4 14.8 Liver size, cm 2 2 4 4 1 2 1 2 2 Spleen size, cm 2 1 out out 0 3 0 1 out Pre-transplant transfusions, units 73 55 470 400 41 32 140 29 250 Serum ferritin, ng/ml 1161 1062 2455 2702 694 2083 1047 1493 1522 Chronic active hepatitis yes no yes yes — no no no no Liver fibrosis, grade 1 0 3 2 0 1 1 0 1 Interval BMT1–BMT2, days 504 1068 1617 3794 714 834 1806 669 1624 Pre-conditioning cytoreductive-immunosuppressive yes yes yes yes yes yes yes yes yes therapy Conditioning regimen BuCy BuCy BuCy BuCy BuCy BuCy BuCy BuCy BuCy GVHD prophylaxis CSMP CSMP CSMP CSMP CSMP CSMP CSMP CSMP CSMP MTX MTX MTX MTX MTX MTX MTX MTX MTX Acute GVHD, grade 0 0 0 0 0 0 0 0 0 Day of thalassemia recurrence 21 — 39 88 — — 23 — — Outcome A&W A&W A&W A&W A&W Died A&W A&W A&W

CSMP = CsA + methylprednisolone (MP); CSMPMTX = CsA + MP + MTX; — = liver biopsy not done. Second BMT for graft failure in thalassemia D Gaziev et al 1304 Therapy-related complications Discussion

None of the patients who received BMT2 experienced Second bone marrow transplantation for malignant and veno-occlusive disease. The incidence of infectious compli- non-malignant diseases is associated with a high morbidity cations was higher in patients after BMT2 than after BMT1, and mortality.7,17 While for patients who have developed but this difference was not statistically significant. Twelve graft failure with an aplastic marrow unresponsive to cyto- patients developed bacteremia by Gram-positive and Gram- kine therapy a second transplant is the only salvage treat- negative species after BMT1, while 18 patients developed ment, the choice of a BMT2 for patients with thalassemia bacterial (four patients), bacterial and/or fungal (seven recurrence is always a difficult decision. Graft failure sec- patients), bacterial-fungal and/or CMV (seven patients) ondary to rejection is a common obstacle to successful mar- infections after BMT2. Five patients (15.6%) developed row transplantation, especially in class 2 and class 3 moderate to severe hemorrhagic cystitis, one patient moder- patients. However, these are patients who have developed ate mucositis, and one patient gastro-intestinal hemorrhage progressive and significant organ damage while receiving after BMT2, while only three patients experienced therapy- conventional treatment. Thus, even a low number of cured related toxicity following BMT1 (moderate mucositis, patients may justify second transplants, as the survival reversible liver failure and moderate to severe hemorrhagic expectancies of such patients are poor in the absence of cystitis, respectively). All patients recovered from marrow transplantation.18 hemorrhagic cystitis with medical (four patients) and Our data showed that the initial engraftment rate for surgical (one patient) management. entire group of patients after BMT2 was lower than it was after BMT1 (67.7% vs 87.5%; P = 0.05), perhaps due to sensitization to donor antigens after BMT1.19 Obviously, Therapy-related death the pattern of graft failure after BMT2 was different from Overall, 11 patients (34%) died after BMT2. Causes of that observed after BMT1 with a prevalence of primary death are shown in Table 4. Two of three patients with graft failure following BMT2. thalassemia recurrence died at home at 215 and 1371 days Our experience of second transplants demonstrates that after BMT2 while receiving conventional treatment (causes a proportion of patients (33%) receiving BMT2 may be unknown). The third patient died of pneumonitis. Three cured. We identified two groups of patients according to patients with primary graft failure died of infectious com- the outcome of graft failure after BMT1: with thalassemia plications while suffering prolonged aplasia. Five patients recurrence and with an aplastic marrow. The probability of died with sustained engraftment. Thus, the transplant- event-free survival was better in the second than in the first related mortality was 28% (nine of 32 patients) in this group of patients (41% vs 29%). In the present study only series. The late death in the patient with a functioning the occurrence of graft failure later than 60 days after graft occurred at day 248 after transplant secondary to BMT1 (secondary late failure) had a positive influence on Pneumocystis carinii infection. survival following BMT2. Survival advantage after BMT2 given for late graft failure has also been demonstrated in patients with aplastic anemia.20 We found that patients who Reinfusion of autologous bone marrow developed graft failure within 60 days after BMT1 and who did not have residual marrow cells had a high incidence of Six patients, three with primary and two with secondary graft failure following BMT2. early graft failure, received autologous bone marrow stored Our data showed that graft failure after transplant may before BMT2. One of these patients had spontaneous recov- be characterized by either the presence or absence of donor ery of granulocytes, while in the remaining five patients the marrow cells. It has been suggested that late graft failure use of G-CSF accelerated granulocyte recovery. Autolog- occurring after initially successful engraftment is presum- ous back-up was given to accelerate platelet recovery. Only ably an immune-mediated phenomenon and may be asso- one of six patients became platelet independent on day 46 ciated with the discontinuation of CsA in some patients.21 after the autologous back-up, while the remaining five were In our series all but two patients who developed secondary platelet transfusion-dependent for 3–5 months after late graft failure were tapering their CsA therapy. The reinfusion. remaining two patients had recently discontinued CsA ther- apy when they developed marrow aplasia. None of these patients responded to high-dose steroids. In these patients the cause of the development of aplasia is probably Table 4 Causes of death immune-mediated, although other mechanisms can not be excluded. It is of interest to note that patients with residual Cause n donor marrow cells before BMT2 had 100% sustained engraftment, probably due to donor–host tolerance.22 Fungal infection (Candida parapsilosis, 2; Aspergillosis, 2) 4 Despite the high success in achieving engraftment, event- Acute respiratory distress syndrome 1 free survival in these patients was low due to high Liver failure plus cytomegalovirus/fungal infection 1 Pneumocytis carinii 1 transplant-related mortality. GVHD 2 The success of second BMT depends in part on increased Unknown (sudden death) 2 immune suppression in the preparative regimen prior to BMT2 compared to that used for the first transplant. In the Second BMT for graft failure in thalassemia D Gaziev et al 1305 present study most patients received TLI and/or ALG in their exemplary care of these patients. This work has been sup- addition to BUCY or CY. Storb et al23 have reported that ported by the Berloni Foundation against Thalassemia, Pesaro and adding ATG to cyclophosphamide as conditioning for a by the Italian Association against Leukemia, Pesaro Section. second transplant for aplastic anemia increased successful engraftment. Neither the dose of CY nor the use of TLI or ALG had influence on survival or rejection after BMT2 in References our study. The ideal conditioning regimen for BMT2 for thalassemia is still to be determined. We have recently 1 Thomas ED, Bukner CD, Sanders JE et al. Marrow transplan- adopted a new preparative regimen for class 3 patients tation for thalassemia. Lancet 1982; 2: 227–228. and patients receiving a second transplant, with some 2 Lucarelli G, Polchi P, Galimberti M et al. Marrow transplan- encouraging results. tation for thalassemia following busulfan and cyclophospham- An increased incidence of acute and chronic GVHD after ide. Lancet 1985; 1: 1355–1357. second BMT has been reported.24 These data differ from 3 Lucarelli G, Galimberti M, Polchi P et al. Bone marrow trans- the current study in which the rates of acute and chronic plantation in patients with thalassemia. New Engl J Med 1990; GVHD were low. This low incidence of acute GVHD may 332: 417–421. be related to the combined GVHD prophylaxis given to 4 Lucarelli G, Giardini C, Baronciani D. Bone marrow trans- plantation in thalassemia. Semin Hematol 1995; 4: 297–303. most of the patients in our series. On the other hand, analy- 5 Lucarelli G, Clift RA, Galimberti M et al. Marrow transplan- sis of the incidence of GVHD is made difficult by the high tation for patients with thalassemia. Results in class 3 patients. rate of primary graft failure after BMT2. Blood 1996; 80: 2082–2088. Moderate to severe hemorrhagic cystitis was the most 6 Lucarelli G, Clift RA, Galimberti M et al. Bone marrow trans- frequent transplant-related toxic complication observed plantation in adult thalassemic patients. Blood 1999; 4: after BMT2, probably due to re-exposure of patients to 1164–1167. cyclophosphamide. In the current study the most frequent 7 Champlin Re, Horowitz MA, van Bekkum DW et al. Graft cause of death was infection, mainly fungal, as has been failure following bone marrow transplantation for severe observed in other studies.24 Fifty-five percent of deaths aplastic anemia: risk factors and treatment results. Blood 1989; occurred before 100 days after BMT2. It is important to 2: 606–613. 8 Nemunaitis J, Singer JW, Buckner CD et al. Use of recombi- note that, the probability of death was higher in patients nant human granulocyte–macrophage colony-stimulating fac- with aplasia compared to patients with thalassemia recur- tor in graft failure after bone marrow transplantation. Blood rence. This reflects the severity of this form of graft failure 1990; 1: 245–253. associated with prolonged neutropenia and a risk of 9 Davis SM, Weisdorf DJ, Haake RJ et al. Second infusion of developing infections which require early intervention. bone marrow for treatment of graft failure after allogeneic The use of cytokines has improved the outcome in bone marrow transplantation. Bone Marrow Transplant 1994; patients with graft failure accompanied by neutropenia.8,9 14: 73–77. An alternative therapy to accelerate myeloid recovery 10 Di Bartolomeo P, Girolamo G, Angrilli F et al. Second mar- might be reinfusion of stored autologous bone marrow row transplant in patients with thalassemia. Bone Marrow before BMT2.25 As our patients had good myeloid recov- Transplant 1993; 12 (Suppl. 1): 78–80. 11 Lin HP, Chan LC, Lam SK et al. Bone marrow transplantation ery, we used autologous reinfusion mainly to accelerate pla- for thalassemia. The experience of Malaysia. Bone Marrow telet recovery. Despite autologous back-up most patients Transplant 1997; 19 (Suppl. 2): 74–77. required considerable time to achieve platelet independence 12 Polchi P, Galimberti M, Lucarelli G et al. Second transplant in this series. Taking into consideration the fact that the in thalassemia. Bone Marrow Transplant 1997; 19 (Suppl. 2): incidence of primary graft failure is particularly high after 83–86. BMT2 and not all patients with graft failure respond to 13 Ramsay NKC, Trewan K, Nesbit E et al. Total lymphoid cytokines, we recommend that autologous bone marrow irradiation and cyclophosphamide as preparation for bone should be harvested and cryopreserved prior to second marrow transplantation in severe aplastic anemia. Blood 1980; transplant. 55: 344–346. These data suggest that second bone marrow transplan- 14 Thomas ED, Storb R, Clift RA et al. Bone marrow transplan- tation. New Engl J Med 1975; 292: 895–902. tation can cure a proportion of patients with a poor survival 15 Kaplan EL, Meier P. Non-parametric estimation from incom- expectancy on conventional treatment. The high rate of plete observations. J Am Stat Assoc 1958; 53: 457–481. graft failure following BMT2 demonstrates that the con- 16 Siegel J (ed). Statistix. Version 4.0 User’s Manual. Analytical ditioning regimens used probably did not provide sufficient Software. Tallahassee: FL, 1992. immunosuppression for successful graft in nearly half of 17 Mrsic M, Horowitz MM, Atkinson K et al. Second HLA- our patients. Results of the new conditioning regimen are identical sibling transplant for leukemia recurrence. Bone encouraging in terms of overall survival and transplant- Marrow Transplant 1992; 9: 269–275. related toxicity and a subsequent increase of immuno- 18 Nathan DG, Oski FA. Hematology of Infancy and Childhood, suppression in this preparative regimen may be effective in vol 1 (ed 4). Saunders: Philadelphia, 1993, pp 844–845. obtaining a higher cure rate. 19 Fleischhauer K, Kernan NA, O’Relly RJ et al. Bone marrow allograft rejection by T lymphocyates recognizing a single amino acid difference in HLA-B44. New Engl J Med 1990; Acknowledgements 323: 1818–1822. 20 McCann SR, Bacigalupo A, Glukman E et al. Graft rejection We thank Ms Aileen Law, a member of our nursing staff, for and second bone marrow transplants for acquired aplastic ane- linguistic assistance and all the members of our BMT Team for mia: a report from Aplastic Anemia Working Party of the Eur- Second BMT for graft failure in thalassemia D Gaziev et al 1306 opean Bone Marrow Transplant Group. Bone Marrow Trans- plants in patients with aplastic anemia rejecting the first graft: plant 1994; 13: 233–237. use of a conditioning regimen including cyclophosphamide 21 Hows J, Palmer S, Gordon-Smith EC. Cyclosporine and stable and antithymocyte globulin. Blood 1987; 1: 116–121. graft failure following bone marrow transplantation for severe 24 Bolger GB, Sullivan KM, Storb R et al. Second marrow aplastic anemia. Br J Haematol 1985; 60: 611–617. infusion for poor graft function after allogeneic marrow trans- 22 Or R, Kapelushnik J, Naparstek E et al. Second transplantation plantation. Bone Marrow Transplant 1986; 1: 21–30. using allogeneic peripheral blood stem cells in a ␤-thalassemia 25 Mehta J, Powles R, Singhal S et al. Outcome of autologous major patient featuring stable mixed chimerism. Br J Haema- rescue after failed engraftment of allogeneic marrow. Bone tol 1996; 94: 285–287. Marrow Transplant 1996; 17: 213–217. 23 Storb R, Weiden PL, Sullivan KM et al. Second marrow trans-