Graft-versus-leukaemia The graft-versus-leukaemia effect in ... - Nature

Bone Marrow Transplantation (2002) 30, 761–768  2002 Nature Publishing Group All rights reserved 0268–3369/02 $25.00 www.nature.com/bmt

Graft-versus-leukaemia The graft-versus-leukaemia effect in haematopoietic stem cell transplantation using unrelated donors M Remberger1,2, J Mattsson1,2, P Hentschke1,2, J Aschan2,3, L Barkholt2, J Svennilson2,4, P Ljungman3 and O Ringde´n1,2 1

Department of Clinical Immunology, Huddinge University Hospital, Stockholm, Sweden; 2Centre for Allogeneic Stem Cell Transplantation, Huddinge University Hospital, Stockholm, Sweden; 3Department of Haematology, Huddinge University Hospital, Stockholm, Sweden; and 4Department of Paediatrics, Huddinge University Hospital, Stockholm, Sweden

Summary: We studied the graft-versus-leukaemia (GVL) effect in 185 patients with haematological malignancies who underwent unrelated donor haematopoietic stem cell transplantation (HSCT) at Huddinge University Hospital between May 1991 and June 2001. Ninety-five were in first CR/CP and 90 in later stages. Most (86%) of them had a HLA-A, -B and -DR␤1 matched donor. Conditioning usually consisted of total body irradiation and cyclophosphamide, and GVHD prophylaxis of cyclosporine and methotrexate. In the multivariate riskfactor analysis of relapse, we found that disease stage beyond CR1/CP1 (P = 0.02), acute GVHD 0–I (P = 0.02), absence of chronic GVHD (P = 0.02) and ALL (P = 0.02) were independent risk factors for relapse. The incidence of relapse in those with acute GVHD grade II was 18% vs 46% in those with no or grade I (P = 0.04). In patients with or without chronic GVHD, the incidences of relapse were 32% and 48%, respectively (P ⬍ 0.01). The best RFS was seen in patients with chronic GVHD. No difference in RFS was seen in patients with no, mild or moderate acute GVHD. Risk factors for relapse after HSCT with unrelated donors were: acute lymphoblastic leukaemia, disease stage beyond CR1/CP1, absence of chronic GVHD and no, or mild acute GVHD. Overall and relapse-free survival were not improved by the occurrence of acute GVHD. Bone Marrow Transplantation (2002) 30, 761–768. doi:10.1038/sj.bmt.1703735 Keywords: HSCT; relapse; leukaemia; GVHD; GVL

Major complications after haematopoietic stem cell transplantation (HSCT) are acute and chronic graft-versus-host disease (GVHD), infections and recurrence of leukaemia.1 Leukaemia relapse occurs in 10–80% of all patients transplanted for a malignant disease, depending on disease,2 dis-

ease stage at transplant,3 donor type,4 occurrence of GVHD,2,5,6 conditioning2 and immune-suppressive regimen.2,7,8 It is well known that patients beyond first complete remission (CR)/chronic phase (CP) have a higher risk of relapse than those in CR1/CP1.3 T cell depletion of the graft also increases risk of relapse, especially in patients with CML.7 GVHD has a graft-versus-leukaemia effect (GVL), since patients with mild acute or chronic GVHD have a significantly lower risk of leukaemia relapse after HSCT, than do those without GVHD.5,6,9 Chronic GVHD seems to have the strongest anti-leukaemia effect,5,10 at least in recipients of sibling grafts. Risk factors for relapse and GVL effect have been extensively analysed when using sibling donors. However, the GVL effect has not been studied in unrelated donors alone. We evaluated risk factors for relapse in 214 consecutive patients with a malignant disease receiving a HSCT from an unrelated donor at Huddinge University Hospital. We also assessed the GVL effect. Patients and material Patients Two hundred and fourteen patients with a haematological malignancy underwent HSCT from an unrelated donor at Huddinge University Hospital between May 1991 and June 2001. One hundred and eighty-five of these patients relapsed or survived for more than 90 days without relapse and were included in the study. Their diagnoses were AML (44), ALL (56), CML (64) and other malignancies (21). Ninety-five were in first CR/CP and 90 in later stages. Most (86%) of them had a HLA-A, -B and -DR␤1 matched donor. Nineteen of them had an HLA allele level mismatched donor (7 HLA-B and 12 HLA-DR) and seven a major HLA class I antigen mismatched donor (three HLA-A and four HLA-B). Patient and donor characteristics are shown in Table 1. Conditioning

Correspondence: M Remberger, Clinical Immunology, F79, Huddinge University Hospital, SE-141 86 Stockholm, Sweden Received 19 March 2002; accepted 8 July 2002

Most patients received cyclophosphamide (CY), 60 mg/kg, on 2 consecutive days, combined with 10 Gy single dose

GVL effect in HSCT with unrelated donors M Remberger et al

762

Table 1

Patient characteristics n

Diagnosis Acute myeloid leukaemia CR1 CR 2–4 Not in remission Acute lymphoid leukaemia CR1 CR 2–4 Not in remission Chronic myeloid leukaemia CP1 CP2 Accelerated phase Lymphoma Myelodysplastic syndrome Other malignancies Recipient age (median, range) Recipient gender (Male/Female) Donor age (median, range) Donor gender (M/F) Matching 6/6 HLA-antigen match Allele level mismatch 1 HLA-antigen mismatch Non-immunized female donor to male recipient Immunized female donor to male recipient AB0 blood group compatibility Identical Minor mismatch Major mismatch

16 (2) 16 (4) 12 (1) 22 (4) 24 (7) 10 (2) 52 (2) 8 4 (2) 3 13 (2) 5 29 (0–60) 99/86 35 (19–56) 113/71 159 19 7 12 17

(86%) (10%) (4%) (6%) (9%)

71 (38%) 52 (28%) 62 (34%)

CR = complete remission. Numbers in brackets shows number of patients with a mismatched donor.

patients with more advanced disease received additional etoposide. All patients received ATG (Thymoglobulin; IMTIX-Sangstat, Lyon, France or ATG-Fresenius; Fresenius AG, Gra¨ felfing, Germany) or OKT-3 (Ortho Biotech, Raritan, NJ, USA) as part of pre-transplant conditioning. The protocols have been described elsewhere.11–15 GVHD prophylaxis Most patients received cyclosporine (CsA) combined with methotrexate (MTX) (Table 2).15 CsA i.v. was started on day ⫺1, the dose ranging from 5 to 10 mg/kg/day, with a higher dose given to small children. On the day of SCT, CsA was lowered to 3–7.5 mg/kg/day. On day +1, or as soon as the patient could take CsA orally, 12.5 mg/kg/day, was given in two doses to adults and a higher dose, 12.5– 20 mg/kg/day, to children. During the first month, blood CsA levels were kept at approximately 200–300 ng/ml.15 CsA was discontinued after 6 months if possible. Seven patients with a major HLA-antigen mismatched donor received T cell-depleted grafts.13 Other immunosuppressive protocols were CsA combined with steroids or mycofenolate mofetil (MMF). Stem-cell source and supportive care Forty-six patients received stem cells from peripheral blood (PBSC) and 139 bone marrow.16 G-CSF was given to 84% of the patients from day 10 after HSCT until ANC ⬎0.5 ⫻ 109/l for 2 consecutive days were reached.17 Relapse

total body irradiation (TBI) (Table 2). Other regimens included busulfan (Bu), 4 mg/kg, on 4 consecutive days combined with CY, or fractionated TBI (4 ⫻ 3 Gy) and CY. Nine patients who received a T cell-depleted graft were conditioned with 120 mg of CY and 3 ⫻ 2 Gy total lymph-node irradiation, followed by TBI 7.5 Gy. Fourteen Table 2

Conditioning and GVHD prophylaxis n (%)

Conditioning TBI 10 Gy/Cy ⫾ VP16 Fractionated TBI 4 ⫻ 3Gy/Cy ⫾ VP16 Busulfan/Cy ATG-Fresenius Thymoglobulin OKT-3 GVHD prophylaxis CsA + MTX T-cell depletion CsA + prednisolone MMF + CsA Stem-cell source PBSC/BM G-CSF post-HSCT Nucleated cell dose (x108/kg) (median, range)

112 38 35 17 127 41

(60) (21) (19) (9) (69) (22)

168 (91) 9 (5) 6 (3) 2 (1) 46/139 (25/75) 155 (84) 3.2 (0.5–38.4)

TBI = total body irradiation; VP16 = etoposide; CY = cyclophosphamide; ATG = anti-thymocyte globuline; CsA = cyclosporine; MTX = methotrexate; MMF = mycophenolate mofetil; PBSC = peripheral blood stem cells; BM = bone marrow; G-CSF = granulocyte colony-stimulating factor. Bone Marrow Transplantation

Patients were, in general, analysed with morphology of bone marrow (BM) aspirates at 3, 6 and 12 months and then yearly after HSCT. A patient with regenerating peripheral blood (PB) values was considered in clinical remission when less than 5% blast cells among 200 nucleated cells were found in a BM sample, on morphological examination. Morphologic leukaemia relapse was defined as at least 30% blast cells in BM aspirates, or presence of extramedullary leukaemia cells. If more than 5% blast cells, but less than 30%, were detected in BM, this was defined as an early relapse. Molecular relapse was defined as increasing levels of BCR/ABL transcript in PB or BM samples in patients with Philadelphia chromosome positive (Ph+) CML or ALL. With other diagnoses, increasing levels of recipient cells in the leukaemia-affected cell lineage were defined as molecular relapse. Diagnosis and treatment of GVHD Both acute and chronic GVHD were diagnosed on the basis of clinical symptoms and/or biopsies (skin, liver, gastrointestinal tract or oral mucosa).18,19 Patients were treated for grade I acute GVHD with prednisolone, starting at 2 mg/kg/day, which was tapered after the initial response.20 In more severe cases, ATG, methylprednisolone, MTX or psoralene and UV light (PUVA) were used. Chronic GVHD was initially treated with CsA and steroids.21 If no response was seen, some patients were given total lymph node


irradiation (TLI),22 PUVA,23 and more recently, extracorporeal PUVA.24 HLA typing All patients and donors were retrospectively HLA class I and class II typed by allele level PCR-SSP.25 Statistics

763

Results

The findings were analysed on 15 August 2001. Time to relapse was determined with the life table method using the log-rank (Mantel–Haenzel) test, taking censored data into account,26 and the Cox regression model was used for uniand multivariate analyses.27 Time to relapse was used to calculate the cumulative incidence curves, and death was a censored observation, unless relapse had occurred. Only patients having a relapse or surviving at least 90 days without relapse after HSCT, and therefore at risk of relapse, were included in the analysis. Twenty-seven potential risk factors for relapse were studied (Table 3). Only factors at ⭐10% levels in the univariate analyses were introduced into the stepwise elimination multivariate analyses. This was done to reduce the numbers of competing factors in the multivariate analysis. When performing multiple comparisons there is always a risk of statistical significance by chance. However, only one multivariate analysis for each event (relapse and relapse-free survival) was done. For this reason no Bonferroni correction was undertaken. A female Table 3

donor was considered as immunized if she had been pregnant or had had a blood transfusion. Patients in CR 1 or CP 1 were considered as early disease, while all others were late disease. An AB0 minor mismatch was a donor who had AB0 antibodies against the patient erythrocytes, while an AB0 major mismatch was a patient who had AB0 antibodies against donor erythrocytes.

GVHD Twenty-eight (15%) patients had no sign of acute GVHD, while 106 (57%) developed grade I, 39 (21%) grade II, 10 (5%) grade III and 2 (1%) grade IV acute GVHD. Seventyfive patients (40%) developed chronic GVHD, 63 (34%) limited and 12 (6%) extensive. Four patients had ‘de novo’ chronic GVHD, while the rest had had acute GVHD previously. Relapse and relapse-free survival Recurrent disease was diagnosed in 56 of the 185 patients. Eleven of these relapses were detected by molecular methods and 45 morphologically. The cumulative incidence of relapse was 38% at 3 years and 41% at 5 years. The cumulative probability of relapse-free survival was 45% at 3 years, and 41% at 5 years.

Univariate Cox regression analysis of risk factors for relapse and relapse-free survival (RFS) in HSCT patients with unrelated donors

Factor

0/1

Relapse RR

Recipient gender Recipient age Chronic myeloid leukaemia Acute lymphoid leukaemia Risk HLA mismatch Nucleated cell dose (⫻108/kg) AB0 minor mismatch AB0 major mismatch Donor gender Donor age Busulfan conditioning Fractionated TBI Thymoglobulin ATG-Fresenius Acute GVHD I–IV Acute GVHD II–IV Acute GVHD III–IV Chronic GVHD Time from diagnosis to HSCT ⬎1 year Female donor to male recipient Immunized female donor to male recipient Stem-cell source CMV reactivation G-CSF post-HSCT Recipient CMV serology Donor CMV serology

Male/Female Continuous No/yes No/yes Low/high No/yes ⬍3.0/⭓3.0 No/yes No/yes Male/Female Continuous No/yes No/yes No/yes No/yes 0/I–IV 0–I/II–IV 0–II/III–IV No/yes No/yes No/yes No/yes BM/PBSC No/yes No/yes ⫺/+ ⫺/+

1.21 0.99 0.44 2.14 2.27 1.58 1.51 1.75 0.73 0.84 1.01 0.54 2.05 1.99 0.52 0.66 0.40 0.56 0.47 1.13 0.99 1.11 1.02 0.58 1.52 1.13 0.83

RFS P 0.48 0.08 0.01 ⬍0.01 ⬍0.01 0.21 0.14 0.04 0.31 0.53 0.76 0.23 0.02 0.03 0.21 0.21 0.03 0.56 ⬍0.01 0.66 0.98 0.81 0.96 ⬍0.05 0.28 0.70 0.54

RR 1.21 1.00 0.64 1.21 1.57 1.42 1.12 1.39 0.84 0.97 1.01 0.86 1.30 1.00 0.67 0.95 1.54 4.14 0.54 0.91 1.00 1.16 1.07 0.94 1.05 1.27 1.01

P 0.92 0.95 0.03 0.31 0.02 0.16 0.55 0.09 0.40 0.89 0.23 0.59 0.25 0.99 0.23 0.86 0.03 ⬍0.001 ⬍0.01 0.61 0.99 0.62 0.76 0.73 0.84 0.28 0.96

Low risk= CR1/CP1; High risk = later stages; AB0 = blood group antigen; HSCT = haematopoietic stem cell transplantation; ATG = anti-thymocyte globulin; BM = bone marrow; PBSC = peripheral blood stem cells; CMV = cytomegalovirus; G-CSF = granulocyte colony-stimulating factor. Bone Marrow Transplantation


764

Table 4a Multivariate Cox regression analysis of risk factors for relapse in 185 HSCT patients with unrelated donors Factor

RR

CI

P

Acute lymphoid leukaemia High risk disease Acute GVHD 0–I Absence of chronic GVHD

1.89 1.99 2.78 1.92

1.10–3.25 1.12–3.54 1.19–6.67 1.10–3.33

0.02 0.02 0.02 0.02

incidence of relapse was 46%, compared to 22% (P = 0.04) in those with grades II–IV acute GVHD (Figure 1c). The incidence of relapse in patients with and without chronic GVHD was 32% and 48%, respectively (P ⬍ 0.01) (Figure 1d). The incidence of relapse increased with the number of risk factors. If no risk factors were present (n = 11), no relapse occurred. If one risk factor was present (n = 37), the cumulative incidence of relapse was 31%, with two risk factors (n = 73), it was 31%, with three (n = 45) it was 54% and with all four present, it was 81%.

Table 4b Multivariate Cox regression analysis of factors for better relapse-free survival in 185 HSCT patients with unrelated donors. Factor

RR

CI

P

Acute GVHD 0–II Chronic GVHD

4.90 2.04

2.47–9.74 1.32–3.23

⬍0.001 0.001

Adjusted for disease and disease stage.

Risk factors for relapse In the Cox regression univariate analysis, eight factors were significant at the ⬍5% levels, and one additional factor at the 5–10% level (Table 3). In the stepwise elimination multivariate analysis, ALL diagnosis, remission beyond CR1/CP1, absence of chronic GVHD and no, or mild acute GVHD were independent risk factors for relapse in this study (Table 4a). Among patients with ALL the 5-year incidence of relapse was 57%, compared to 31% (P ⬍ 0.01) in patients with another diagnosis (Figure 1a). Among patients beyond CR1/CP1, the 5-year incidence of relapse was 50%, compared to 32% (P ⬍ 0.01) in those in CR1/CP1 (Figure 1b). Among patients with no, or mild acute GVHD, the 1.0

Graft-versus-leukaemia effect Among patients with acute GVHD II, the cumulative incidence of relapse was 18%, compared to 46% in all others (P = 0.04). We found no difference in the incidence of relapse between patients with grade I and those without acute GVHD (45% vs 48%). Chronic GVHD was associated with a lower incidence of relapse. The combination of acute GVHD grade II and chronic GVHD was not associated with a lower incidence of relapse than in patients with acute GVHD grade II without chronic GVHD (Figure 2). Patients with isolated acute GVHD grade I showed no difference in relapse incidence compared to those without any GVHD, while those with both acute GVHD grade I and chronic GVHD had a lower incidence (Figure 2). Among patients with grade III–IV acute GVHD (n = 12) only one relapsed. However, only four of these patients survived more than 6 months, and two for more than 1 year. In this study, moderate-to-severe chronic GVHD was not associated with a stronger antileukaemia effect than was mild disease. 1.0

a

0.4

CML 31% n = 64 AML 30% n = 44

0.2

1.0

P = 0.001

Relapse

Relapse

ALL 57% n = 56

0.6

0.0

0.2 0.0

0.4

0.0

0 12 24 36 48 60 72 84 96 108120 132 Months after HSCT

1.0

c

P = 0.003 Early disease 32% n = 95


d

0.8 aGVHD IIIIV n = 12 No aGVHD n = 29 aGVHD I n = 105 aGVHD II n = 39

P < 0.05


Relapse

Relapse

0.4

Late disease 50% n = 60

0.6

0.2

0.8 0.6

b

0.8

0.8

0.6

P < 0.01

No aGVHD 48% n = 109

0.4 0.2 0.0

cGVHD 32% n = 75 0 12 24 36 48 60 72 84 96 108120 132 Months after HSCT

Figure 1 Cumulative incidence of relapse in 185 HSCT patients with unrelated donors according to (a) the diagnosis, (b) disease stage, (c) grade of acute GVHD and (d) chronic GVHD. Bone Marrow Transplantation

1.0

1.0

0.8

0.8

0.6

Relapse-free survival

Relapse


aGVHD I, no cGVHD No GVHD

0.4

aGVHD I+cGVHD aGVHD II+cGVHD aGVHD II, no GVHD

0.2

0.0

765

aGVHD I+cGVHD

0.6

aGVHD II, no cGVHD

aGVHD II+cGVHD no GVHD

0.4

aGVHD I, no cGVHD

0.2 aGVHD IIIIV +/- cGVHD

0

20

40

60 80 100 Months after HSCT

120

140

Figure 2 Cumulative incidence of relapse in 185 HSCT patients with unrelated donors and various combinations of acute and chronic GVHD.

0.0

0

20

40

60 80 Months after HSCT

100

120

140

Figure 3 Cumulative incidence of relapse-free survival in 185 HSCT patients with unrelated donors and various combinations of acute and chronic GVHD.

Survival The overall 5-year survival in the whole group was 46%. Survival rates among patients with acute GVHD grades 0, I, II and III–IV were 57%, 50%, 49% and 12%, respectively. Patients with chronic GVHD showed a significantly better 5-year survival than did those without, 63% vs 45%, P ⬍ 0.01. Transplant-related mortality (TRM) was 23% for patients with no or mild acute GVHD while it was 43% in patients with grade II acute GVHD (P = 0.02). Relapse-free survival In the multivariate risk factor analysis of RFS, we found that chronic GVHD and grades 0–II acute GVHD correlated independently with a better RFS, when adjusted for diagnosis and disease stage (Table 4b). The 5-year RFS of patients with chronic GVHD was 55%, compared to 39% in those without (P ⬍ 0.01). The 5-year RFS in patients with acute GVHD grades 0–II was 46%, compared to 12% in those with more severe forms (P ⬍ 0.001). We found no differences in RFS rates among those without, with mild or with grade II acute GVHD (49%, 46% and 47%, respectively). The cumulative incidence of RFS in patients with chronic GVHD and acute GVHD 0–II (n = 73) was 57%. RFS rates in patients with various combinations of acute and chronic GVHD are shown in Figure 3. DLI treatment Twenty-five of the patients who relapsed were treated with donor lymphocyte infusion (DLI). Seven (28%) of these entered a new remission after the DLI treatment. Two had only a transient remission, while five are still in CR, with 2–72 months follow-up. Response rates for the different diagnoses are given in Table 5. In 11 of the DLI-treated patients, relapse was detected with molecular methods, while 14 had a morphologically detected relapse. Among those with a relapse detected by molecular methods, four responded, and all are still alive in remission, while in those with a morphological relapse three responded, and one is still alive in remission. One of eight patients with acute

leukaemia who relapsed within 1 year after HSCT responded to DLI treatment. Among those who had a later relapse, two of four patients responded. After DLI treatment, 16 patients had no acute GVHD, four grade I, two grade II and three grades III–IV. Chronic GVHD occurred in seven of these patients (Table 5). Among those still alive and in remission (n = 5) after DLI treatment, two did not develop acute GVHD, two had mild and one moderate disease, while chronic GVHD occurred in four. Of those who had only a short remission, one had acute GVHD and one chronic GVHD. On logistic regression analysis, chronic GVHD (OR 20.0, 95% CI 1.96–204, P ⬍ 0.01) was strongly correlated with induction of a new remission after DLI, while acute GVHD was not. Discussion Various factors affecting the incidence of recurrence of leukaemia after HSCT have been reported.2 However, previous studies have mainly been undertaken in patients receiving related donor transplants. These have shown that acute GVHD of any grade, and chronic GVHD, play an important role in lowering the incidence of leukaemia relapse.2,5,6,9 The findings of these studies together with the observation that donor lymphocyte infusions can cure patients with relapse28–31 have been interpreted as indicating that GVHD is associated with a graft-versus-leukaemia (GVL) effect. The present study is the first to evaluate risk factors for relapse and the GVL effect solely in patients receiving transplants from unrelated donors. We found that patients with ALL had a significantly higher incidence of relapse than those having other diagnoses. Many of them (61%) were beyond first remission. This was expected because results with chemotherapy in children with ALL are very good. This means that only the most severe cases are transplanted, leading, not surprisingly to a high relapse rate. However, most patients with AML (64%) were also beyond first CR. Furthermore, an ALL diagnosis was a risk factor for relapse independent of disBone Marrow Transplantation


766

Table 5

Response rates after donor-lymphocyte infusion (DLI) treatment of relapse after haematopoietic stem cell transplantation (HSCT)

Diagnosis

AML ALL CML Others Chronic GVHD No chronic GVHD

Relapse

Cytogenitic Morphologic Cytogenetic Morphologic Cytogenetic Morphologic Cytogenetic Morphologic

n Response

Alive in CR/CP

Alive with disease

Acute Chronic Follow-up for patients alive GVHD GVHD in CR/CP (in months)

1 2 4 5 5 6 1 1

0 0 2 1 2 2 0 0

0 0 2 0 2 1 0 0

0 0 1 0 3 1 0 1

0 1 3 0 1 4 0 0

0 0 2 2 2 1 0 0

— — +2, +15 — +38, +70 +72 — —

7 18

5 2

4 1

1 5

3 5

7 0

+2, +15, +70, +72 +38

AML = acute myeloid leukaemia; ALL = acute lymphoid leukaemia; CML = chronic lymphoid leukaemia. Numbers of patients are presented. Follow up for patients alive and in CR/CP are given in months from the patient entered the remission.

ease stage on multivariate analysis. Many of the patients with ALL considered to be in CR have recently been shown to harbour high levels of minimal residual disease (MRD), which was associated with an increased relapse rate.32 These results may indicate that ALL is less responsive to a GVL effect than are other haematological malignancies. All patients with haematological malignancies given a HSCT from an unrelated donor could be considered at high risk for relapse. Patients in first CR have high-risk criteria, such as chromosomal aberrations, high leukocyte counts at diagnosis or difficulty in achieving remission.33 In this study, however, we did not evaluate these criteria. Many studies have shown that patients beyond first CR/CP are at high risk of relapse after HSCT,3 which was confirmed by our study. Here, this effect was most obvious in patients with AML, where none of those in CR1 had a relapse (n = 16) vs 56% of those with later disease stages (n = 28). Among patients with ALL, the corresponding figures were 41% vs 64%. It is also well known that relapse rates are lower in patients with GVHD.2,5,6,9 This GVL effect is thought to be strongest in chronic GVHD,5,10 while acute GVHD of any grade has a weaker anti-leukaemia effect.5,10,34 We found no difference in relapse rates between patients without acute GVHD and those with mild disease. This may be explained by a suggested GVL effect even in the absence of GVHD.35 In the multivariate risk factor analysis in our study, no, or mild acute GVHD was a risk factor for relapse (P = 0.02), but absence of acute GVHD compared to any grade was not. Chronic GVHD was also correlated with a lower incidence of relapse (P = 0.02). In the present study, acute GVHD grade II had a strong GVL effect. We failed to find a stronger GVL effect in patients with grades III– IV acute GVHD. The reason for this is that very few (2/12) of these patients survived more than 1 year after HSCT. The lowest relapse rate was found in patients with acute GVHD grade II and chronic GVHD, but this was not significantly better than among patients with acute GVHD grade II without chronic GVHD (Figure 2). Since patients with haematological malignancies given a HSCT from an unrelated donor could be considered as high risk, they probably need a stronger GVL effect to eradicate their disease. This is probably why we failed to find a GVL effect from Bone Marrow Transplantation

mild acute GVHD in these patients. In contrast, among patients with leukaemia given marrow from an HLA-identical sibling, those with acute GVHD grade I had the best RFS.36 For such a transplant, patients with lower risk of relapse are also considered and therefore even mild GVHD had a GVL effect. In accordance with the risk factor analysis for relapse, we found the best RFS rates in patients with chronic GVHD and acute GVHD grades 0–II. However, long-term survival was not different from those without GVHD (Figure 3). The lower relapse incidence in patients with acute GVHD grades II–IV was diminished by a higher transplant-related mortality (TRM) among these patients. As a result we did not find any difference in RFS rates in patients without, or with mild or moderate acute GVHD (49%, 46% and 47%), while grades III–IV gave a very poor outcome (Figure 3). This indicates that GVHD is something to be avoided. Response rates to DLI have been shown to depend on time of relapse, disease, induction of GVHD and grade of disease at treatment.37 The best results have been found in patients with CML, while those with ALL have very low response rates29,38 suggesting that the GVL effect is weaker in this group. However, in the present study three of nine patients with ALL responded to DLI treatment, which was comparable to the response rate in patients with CML (4/11) (Table 5). Our data suggest that patients with ALL relapsing after HSCT may be salvaged by cell therapy. This is in line with observations by Slavin et al.39 but in contrast to other reports.29,38 It has previously been shown that response rates in patients with acute leukaemia who relapse early (within 1 year after HSCT) are far worse than if the relapse occurs later.37 This was confirmed in this study. However, the numbers are small and many more patients are needed for a conclusion on this point. DLI treatment induced remission in seven of 25 (28%) treated patients. Among those, we found a correlation between chronic GVHD and a new remission (P ⬍ 0.01). Five of the seven patients with chronic GVHD entered a new remission (Table 5). Early detection and treatment of recurrent disease after HSCT has been shown to positively affect results.37 With sensitive molecular methods recurrent disease may be detected several weeks or months prior to morphological relapse.40,41 In the present study, four of the five patients


with a long-lasting remission received their DLI treatment after a relapse detected by molecular means. This may indicate that response rates after DLI treatment in acute leukaemia may improve with early detection of recurrent disease. We conclude that using unrelated donors, the strongest anti-leukaemia effect was seen in acute GVHD grade II. It may be risky to attempt and induce grade II acute GVHD in recipients of unrelated grafts to have the strongest possible GVL effect. This may not be necessary because overall and RFS were as good in patients with no, or mild GVHD only (Figure 3). After DLI treatment from unrelated donors, chronic GVHD was correlated with a high rate of remission induction. Acknowledgements We thank the staff at the Centre for Allogeneic Stem-cell Transplantation, Departments of Haematology and Paediatrics for excellent patient care. This study was supported by grants from Dagmar Ferbs Memorial foundation, Svenska La¨ karefo¨ rbundet (2000-02-0553), the Swedish Cancer Society (0070-B97-11XBC), the Children’s Cancer Foundation (1997/073, 1998/002), the Swedish Medical Research Council (K98-06X-05971-18B) and the Tobias Foundation (B11/98).

References 1 Sullivan KM, Weiden PL, Storb R et al. Influence of acute and chronic graft-versus-host disease on relapse and survival after bone marrow transplantation from HLA-identical siblings as treatment of acute and chronic leukemia. Blood 1989; 73: 1720–1728. 2 Carlens S, Ringde´ n O, Aschan J et al. Risk factors in bone marrow transplant recipients with leukaemia. Increased relapse risk in patients treated with ciprofloxacin for gut decontamination. Clin Transplant 1998; 12: 84–92. 3 Ringde´ n O, Zwaan F, Hermans J, Gratwohl A, for the Leukemia Working Party of the European Group for Bone Marrow Transplantation. European experience of bone marrow transplantation for leukemia. Transplant Proc 1987; 19: 2600. 4 Perreault C, Decary F, Brochu S et al. Minor histocompatibility antigens. Blood 1990; 76: 1269–1280. 5 Ringde´ n O, Labopin M, Gluckman E et al. for the Acute Leukemia Working Party of the European Group for Blood and Marrow Transplantation. Graft-versus-leukemia effect in allogeneic marrow transplant recipients with acute leukemia is maintained using cyclosporin A combined with methotrexate as prophylaxis. Bone Marrow Transplant 1996; 18: 921–929. 6 Horowitz M, Gale R, Sondel P et al. Graft-versus-leukemia reactions following bone marrow transplantation in humans. Blood 1989; 75: 555–562. 7 Goldman J, Gale R, Horowitz M et al. Bone marrow transplantation for chronic myelogenous leukemia in chronic phase: increased risk of relapse associated with T-cell depletion. Ann Int Med 1988; 108: 806–814. 8 Ringde´ n O, Horowitz M, Sondel P. Methotrexate, cyclosporine or both to prevent graft-versus-host disease after HLAidentical sibling bone marrow transplants for early leukemia. Blood 1993; 18: 1094–1101. 9 Weiden P, Flournoy N, Thomas E et al. Antileukemic effect of graft-versus-host disease in human recipients of allogenic marrow grafts. New Engl J Med 1979; 300: 1068–1073.

10 Weiden PL, Sullivan KM, Flournoy N et al. Antileukemic effect of chronic graft-versus-host disease: contribution to improved survival after allogeneic marrow transplantation. New Engl J Med 1981; 304: 1529–1533. 11 Ringde´ n O, Baryd I, Johansson B et al. Increased mortality by septicemia, interstitial pneumonitis and pulmonary fibrosis among bone marrow transplant recipients receiving an increased mean dose rate of total irradiation. Acta Radiol Oncol 1983; 22: 423–428. 12 Ringde´ n O, Ruutu T, Remberger M et al. for the Nordic Group of Bone Marrow Transplantation. A randomized trial comparing busulphan with total body irradiation as conditioning in allogenic marrow transplant recipient with leukemia. Blood 1994; 83: 2723–2730. 13 Ringde´ n O, Remberger M, Aschan J et al. Long-term followup of a randomized trial comparing T-cell-depletion with a combination of methotrexate and cyclosporin in adult leukemic marrow transplant recipients. Transplantation 1994; 58: 887–891. 14 Remberger M, Svahn B-M, Hentschke P et al. Effect on cytokine release and graft-versus-host disease of different anti-T cell antibodies during conditioning for unrelated haematopoietic stem cell transplantation. Bone Marrow Transplant 1999; 24: 823–830. 15 Ringde´ n O, Remberger M, Persson U et al. Similar incidence of graft-versus-host disease using HLA-A, -B and -DR identical unrelated bone marrow donors as with HLA-identical siblings. Bone Marrow Transplant 1995; 15: 619–625. 16 Remberger M, Ringde´ n O, Blau I et al. No difference in graftversus-host disease, relapse and survival comparing peripheral stem cells with bone marrow using unrelated donors. Blood 2001; 98: 1739–1745. ¨ hman S et al. A prospective ran17 Ha¨ gglund H, Ringde´ n O, O domized trial of Filgrastim (r-metHuG-CSF) given at different times after unrelated bone marrow transplantation. Bone Marrow Transplant 1998; 24: 831–836. 18 Thomas E, Storb R, Clift R et al. Bone marrow transplantation I and II. New Engl J Med 1975; 292: 832–843, 895–902. 19 Shulman H, Sullivan K, Weiden P et al. Chronic graft-versushost syndrome in man: a clinicopathological study of 20 longterm Seattle patients. Am J Med 1980; 69: 204–217. 20 Ringde´ n O, Ba˚ ryd I, Gahrton G et al. Early treatment and prophylaxis of graft-versus-host disease using prednisolone. Bone Marrow Transplant 1981; 2: 171–177. 21 Sullivan KM, Witherspoon RP, Storb R et al. Alternating-day cyclosporine and prednisone for treatment of high-risk chronic graft-v-host disease. Blood 1988; 72: 555–561. 22 Socie G, Devergie A, Cosset JM et al. Low-dose (1 Gy) total lymphoid irradiation for extensive, drug-resistant chronic graft-versus-host disease. Transplantation 1990; 49: 657–658. 23 Hymes SR, Morison WL, Farmer ER et al. Methoxsalen and ultraviolet A radiation in treatment of chronic cutaneous graftversus-host reaction. J Am Acad Dermatol 1985; 12: 30–37. 24 Besnier DP, Chabannes D, Mahe B et al. Treatment of graftversus-host disease by extracorporeal photochemotherapy: a pilot study (see comments). Transplantation 1997; 64: 49–54. 25 Olerup O, Zetterqvist H. HLA-DR typing by PCR amplification with sequence-specific primers (PCR-SSP) in 2 hours. An alternative to serological DR typing in clinical practice including donor-recipient matching in cadaveric transplantation. Tissue Ant 1992; 39: 225–235. 26 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. 27 Cox D. Regression models and life-tables. JR Stat Soc (Series B) 1972; 34: 187–220.

767

Bone Marrow Transplantation


768

28 Kolb HJ. Donor leukocyte transfusions for treatment of leukemic relapse after bone marrow transplantation. EBMT Immunology and Chronic Leukemia Working Parties. Vox Sang 1998; 74: 321–329. 29 Kolb H-J, Schattenberg A, Goldman J et al. for the European Group for Blood and Marrow Transplantation Working Party Chronic Leukemia. Graft-versus-leukemia effect of donor lymphocyte transfusions in marrow-grafted patients. Blood 1995; 86: 2041–2050. 30 Dazzi F, Szydlo RM, Craddock C et al. Comparison of singledose and escalating-dose regimens of donor lymphocyte infusion for relapse after allografting for chronic myeloid leukemia. Blood 2000; 95: 67–71. 31 Mackinnon S, Papadopoulos EB, Carabasi MH et al. Adoptive immunotherapy evaluating escalating doses of donor leukocytes for relapse of chronic myeloid leukemia after bone marrow transplantation: separation of graft-versus-leukemia responses from graft-versus-host disease. Blood 1995; 86: 1261–1268. 32 Uzunel M, Mattsson J, Jaksch M et al. The significance of graft-versus-host disease and pretransplant minimal residual disease status to outcome after allogeneic stem cell transplantation in patients with acute lymphoblastic leukemia. Blood 2001; 98: 1982–1984. 33 Gale RP, Horowitz MM, Weiner RS et al. Impact of cytogenetic abnormalities on outcome of bone marrow transplants in acute myelogenous leukemia in first remission. Bone Marrow Transplant 1995; 16: 203–208. 34 Horowitz MM, Gale RP, Sondel PM et al. Graft-versus-leukemia reactions after bone marrow transplantation. Blood 1990; 75: 555–562. 35 Ringde´ n O, Labopin M, Gorin NC et al. Is there a graft-ver-

Bone Marrow Transplantation

36

37

38

39

40

41

sus-leukaemia effect in the absence of graft-versus-host disease in patients undergoing bone marrow transplantation for acute leukaemia? Br J Haematol 2000; 111: 1130–1137. Ringde´ n O, Hermans J, Labopin M et al. for the Acute and Chronic Leukaemia Working Parties of the European Group for Blood and Marrow Transplantation (EBMT). The highest leukaemia-free survival after allogeneic bone marrow transplantation is seen in patients with grade I acute graft-versushost disease. Leuk Lymphoma 1996; 24: 71–79. Carlens S, Remberger M, Aschan J, Ringde´ n O. The role of disease stage in the response to donor lymphocyte infusions as treatment for leukemic relapse. Biol Blood Marrow Transplant 2001; 7: 31–38. Collins RH Jr, Shpilberg O, Drobyski WR et al. Donor leukocyte infusions in 140 patients with relapsed malignancy after allogeneic bone marrow transplantation (see comments). J Clin Oncol 1997; 15: 433–444. Slavin S, Naparstek E, Nagler A et al. Allogeneic cell therapy with donor peripheral blood cells and recombinant human interleukin-2 to treat leukemia relapse after allogeneic bone marrow transplantation. Blood 1996; 87: 2195–2204. Zetterquist H, Mattsson J, Uzunel M et al. Mixed chimerism in the B cell lineage is a rapid and sensitive indicator of minimal residual disease in bone marrow transplant recipients with pre-B cell acute lymphoblastic leukemia. Bone Marrow Transplant 2000; 25: 843–851. Mattsson J, Uzunel M, Tammik L et al. Leukemia lineagespecific chimerism analysis is a sensitive predictor of relapse in patients with acute myeloid leukemia and myelodysplastic syndrome after allogeneic stem cell transplantation. Leukemia 2001; 15: 1976–1985.