Relapse The impact of donor gender on outcome of ... - Nature

8 downloads 0 Views 208KB Size Report
Feb 7, 2005 - European Group for Blood and Marrow Transplantation. (EBMT). The best ..... Also, a separate comparison of the relapse risk in F-M and M-M in.
Bone Marrow Transplantation (2005) 35, 609–617 & 2005 Nature Publishing Group All rights reserved 0268-3369/05 $30.00

www.nature.com/bmt

Relapse The impact of donor gender on outcome of allogeneic hematopoietic stem cell transplantation for multiple myeloma: reduced relapse risk in female to male transplants G Gahrton1, S Iacobelli2, J Apperley3, G Bandini4, B Bjo¨rkstrand1, J Blade´5, JM Boiron6, M Cavo4, J Cornelissen7, P Corradini8, N Kro¨ger9, P Ljungman1, M Michallet10, NH Russell11, D Samson3, A Schattenberg12, B Sirohi13, LF Verdonck14, L Volin15, A Zander9 and D Niederwieser16 1

Departments of Medicine and Hematology, Karolinska Institute and Karolinska University Hospital, Huddinge, Stockholm, Sweden; Leiden University Medical Center, Netherlands; 3Department of Haematology, Faculty of Medicine, Imperial College, London, UK; 4 Institute of Hematology ‘L. and A. Sera`gnoli’, University of Bologna, S Orsola-Malpighi Hospital, Bologna, Italy; 5Hematology, Hospital Clı´nic, Villarroel 170, Barcelona, Spain; 6CHU Bordeaux, Department of Hematology, Hopital Haut-Leveque, Bordeaux, France; 7Haematology Daniel den Hoed Cancer Centre (AZR), Rotterdam, The Netherlands; 8Istituto Nazionale dei Tumori, Ematologia-Unita’ di trapianto di Midollo osseo, via Venezian 1, Milano, Italy; 9University Hospital Eppendorf, Bone Marrow Transplantation, Martinistrasse 52, Hamburg, Germany; 10Hoˆpital Edouard Herriot, Hop. Edouard Herriot Serv. d’hematologie U E4, Lyon Cedex 03, France; 11Department of Haematology, Nottingham City Hospital NHS Trust, Hucknall Road, Nottingham, UK; 12UMC St Radboud, Hematology, Geert Grooteplein Zuid 8, Nijmegen, The Netherlands; 13Royal Marsden, Leukaemia and Myeloma Unit, Downs Rd, Sutton, Surrey, UK; 14AZ Utrecht, Department of Haematology (GO3.647), Utrecht, The Netherlands; 15 Division of Hematology, Helsinki University Central Hospital, Helsinki, Finland; and 16Division of Hematology and Oncology, Department of Internal Medicine, University of Leipzig, Philipp Rosenthalstr. 23-25, Leipzig, Germany 2

Summary: The impact of the donor gender on outcome in HLAidentical sibling donor hematopoietic stem cell transplantation for multiple myeloma was studied in a retrospective registry study of 1312 patients (476 male to male (MM); 334 female to male (F-M); 258 male to female (M-F); 244 female to female (F-F) reported to the European Group for Blood and Marrow Transplantation (EBMT). The best overall survival (OS) from the time of transplantation was found in F-F (median 41 months) with no significant difference between other groups (median 25 months in M-M, 18 months in F-M, 19 months in M-F) despite a significantly higher nonrelapse mortality in F-M. This was due to a significantly lower relapse rate (REL) in F-M compared to all other groups. Before 1994, OS was poorer in F-M than in M-M, which improved to similarity from 1994 onwards (median 29 months in M-M and 25 months in F-M). The reduced REL contributed to this improvement in F-M indicting a gender-specific graft vs myeloma effect. Therefore, a female donor is as good as a male one for male patients, while for female patients gender disparity is a negative factor for outcome.

Correspondence: Dr G Gahrton, Karolinska Institute, Department of Medicine, Karolinska University Hospital, Huddinge, SE 14186 Stockholm, Sweden; E-mail: [email protected] Received 2 September 2004; accepted 9 December 2004 Published online 7 February 2005

Bone Marrow Transplantation (2005) 35, 609–617. doi:10.1038/sj.bmt.1704861 Published online 7 February 2005 Keywords: multiple myeloma; allogeneic stem cell transplantation

Multiple myeloma is a disease characterized by a median overall survival (OS) of about 3 years with conventional chemotherapy and no cures.1–4 Autologous transplantation has improved survival but cures are unlikely to occur.5–7 Allogeneic transplantation has been carried out since the early 1980s.8–10 This treatment modality has been hampered by high transplantation-related mortality,11,12 but occasionally patients have been long-term survivors without signs of disease. Using myeloablative conditioning treatment, the relapse rate (REL) has proven to be lower than with autologous transplantation.11 Sustained molecular remissions are observed in approximately 30% of patients and predict for improved survival.13 Prognostic factor analyses have been performed repeatedly by the EBMT10,12 and other groups, in particular the Seattle group.14,15 Many of the factors of importance for outcome following allogeneic transplantation are the same as for autologous transplantation or conventional chemotherapy. Favorable prognostic factors are low age, low stage at diagnosis, low beta2-microglobulin at diagnosis, few treatment regimens before the transplant, and response to previous chemotherapy. It has been difficult to find specific procedural factors of importance for prognosis. Thus, no conditioning

Donor gender in transplantation of myeloma G Gahrton et al

610

regimen has proved to be superior to the conventional total body irradiation (TBI) þ cyclophosphamide treatment, and no graft-vs-host prevention method has proved to be superior to the classical combination of methotrexate and cyclosporin. A procedural factor that is specific to allogeneic transplants is the donor. It is well known that HLA compatibility is a favorable prognostic factor and a prerequisite for successful hematopoietic stem cell transplantation in most disorders.16 Also, in multiple myeloma, sibling donor transplants have a better outcome than transplants from unrelated donors,17 although only a few such transplants have been performed in myeloma and the results are therefore uncertain. Previous studies have indicated that the gender of the sibling donor is of importance, not only in transplantation of multiple myeloma patients12,14 but also in other hematological disorders.18–20 In one study of myeloma patients, there was a tendency for better survival in female to female transplants than female to male transplants,12 and in another study a transplant from a male donor to a female patient was an adverse risk factor.14 Various reasons for the disparity in outcome have been discussed, and in some studies there has been an influence only on transplantrelated mortality, while in others there has been a difference in REL. The recent findings of minor histocompatibility antigens encoded by genes on the Y chromosomes and their possible importance in a graft vs tumor effect20–25 was an additional reason to perform the present study. We have now investigated 1312 patients reported to the EBMT on so-called MED-B forms (complete data set) transplanted with HLA-matched sibling donors in an attempt to delineate the importance of donor gender in transplantation of patients with multiple myeloma.

Patients and methods The study population comprised 1312 patients reported from 186 centers to the EBMT registry from 1983 to 2001. In total, there were 502 female and 810 male patients. Patients who had received a transplant from an HLAmatched sibling donor and in whom data had been reported on MED-B forms (complete data set) were included. The patient characteristics of the four groups male donor-male patient (M-M), female donor–male patient (F-M), male donor–female patient (M-F), and female donor–female patient (F-F), are shown in Table 1. A prerequisite for including a patient in the study was the knowledge of the gender of both patient and donor and the outcome of the transplant. As can be seen in Table 1, the proportion of patients in the different known prognostic subgroups for male and female subjects, respectively, was similar. Most important was a similar distribution of patients who had received only one line of pretreatment before transplant as opposed to two or more, and of patients who were in a complete remission (CR) as opposed to being in only partial remission, not responding or in progression. Also, the distribution of procedural factors, that is, source of graft, conditioning regimen and graft-vshost prevention, was relatively similar between the groups. The median follow-up was 36 months, with no significant Bone Marrow Transplantation

difference among the four groups. CR after transplantation was defined for the purpose of this study as the disappearance of abnormal immunoglobulins from serum or light chain from the urine using either conventional electrophoresis or immunofixation, as well as disappearance of apparent myeloma cells from the marrow, as previously described.10,12,26

Statistical analysis In the EBMT Myeloma registry we found, for the period between 1983 and 2001, 1419 patients treated with allogeneic hematopoietic stem cell transplantation from HLA-identical siblings for whom basic information was available. Among these only 3% had missing information on donor gender, and therefore we were able to select 1376 potential patients for the study. Since the information regarding the main prognostic factors was not complete in all patients, it was decided to select cases with complete information on the interval diagnosis-transplant and the use of total body irradiation (TBI), since this reduced the population by less than 5%, and to consider the level ‘missing’ as a further category for the other variables (the treatment of missing values is illustrated below). Therefore, the final number of cases included in the study was 1312. Differences in the distributions of the risk factors in different groups were evaluated by a chi-square test on the appropriate cross tabulations for the discrete variables, and by a Mann–Whitney test for the continuous variables. Survival probabilities for OS and relapse-free survival (RFS) were estimated according to the Kaplan–Meier product limit method and differences among groups were tested by the Log-Rank test. The analysis of REL and nonrelapse mortality (NRM) as competing risks was carried out estimating the cumulative incidence curves by the proper estimator, and testing the differences using the Gray test.27 The CR rate was also estimated in a framework of competing risks, the competing event being death without CR. This explains why in previous studies where the competing risks situation was ignored and the cumulative incidence was estimated as the One-Minus-Survival probability obtained from the Kaplan–Meier curve, higher CR rates were reported for the same kind of patients: in fact, the latter method overestimates the cumulative incidence.28,29 Moreover, due to the structure of the Med-B forms it was not possible to distinguish between cases where CR was achieved but the information was missing and cases where CR was never achieved (in particular, when CR was achieved beyond the first 100 days); therefore, all missing cases were considered as ‘no CR’, which means that the estimated rate is the minimum rate observed in the population, but the real rate could be higher. The effect on the hazard of each of the four major outcomes (OS, RFS, REL, and NRM) of gender combination was estimated via multivariate Cox models to adjust for the main prognostic factors known from the literature (see below). The results are reported in Table 2 as Hazard ratios (HR) with respect to a reference category (indicated in the tables by HR ¼ 1) with 95% confidence interval (CI) and P-values.

Donor gender in transplantation of myeloma G Gahrton et al

611 Table 1

Patient characteristics M-M

F-M

476

334

Total number of patients % eval Age class p45 years 445 years

100

82

Stage at diagnosis I II III

67

Beta2-microglobulin Low (p4 mg/l) High (44 mg/l)

32

Treatment lines 1 line 41 line

63

Interval diagn-transpl o6 months 6–12 months 412 months

100

TBI given No Yes

100

Years of transplant o1994 X1994

0.12

100

81

76

0.35 54 24 18 2 1

60

37

0.27 14 23 63

29 70 30

0.76

0.13 53 48

12 17 71 0.90

% of eval

66

0.77 72 28

57

0.45

57 43

55 45

22 21

24 19

25 18

30 15

48

0.71

52

47

0.16

13 87

14 86

11 89

16 84

58 18 8 3

56 20 5 5

68 14 4 3

55 16 10 3

100

0.38

0.09

100

88

0.30 62 38

83 76 24

0.26 22 78

0.59 66 34

57 43 0.35

100

45

100

68 32

0.52 35 65

63 37 0.33

85 71 29

100

54

50 50

0.60 14 34 52

38 62 0.26

100 54 46

100

100

65 35

0.74 78 23

12 38 50 0.15

53

25 75

100

41 59

70 30

100 79 21

10 36 54 100

100

100

75 25 100

50

87

% eval

57 43

36 64

T-cell depletion No Yes

% of eval

P

57 43

13 41 47

100

244

50 20 26 3 1

66 34

78 23

Stem cells used BM PB

258

81

17 21 62

57

F-F

46 54 0.50

29

M-F

100

62 18 15 3 3 67

52

100

0.23

83

67 33

Previous auto-transpl No Yes

% eval

45 55

11 22 67

Details no CR: PR Stat/no response Progression Other

% of eval

100

54 20 18 5 3

Details 41 line: 2 lines 3 lines

Reduced-intensity conditioning No Yes

% eval

49 51

MM class IgG IgA Light chain Nonsecretory Others

Response before transplantation CR No CR

% of eval

P

100

0.56 73 27

100 23 77

0.05 31 69

The proportion of patients in the prognostic subgroups for male and female subjects respectively, was similar. BM ¼ bone marrow cells; PB ¼ peripheral blood stem cells. % eval ¼ percentage evaluable patients of all patients in the donor–patient group. % of eval ¼ percentage patients of those evaluable in the donor–patient group. P-values are from a chi-square test to compare the distribution of the risk factor with respect to donor gender, separately for male patients and female patients. M-M ¼ male donor-male recipient; F-M ¼ female donor-male recipient; M-F ¼ male donor-female recipient; F-F ¼ female donor - female recipient.

Bone Marrow Transplantation

Donor gender in transplantation of myeloma G Gahrton et al

612 Table 2

Effect of gender combination (donor- recipient) (multivariate analysis) OS

M-M F-M M-F F-F

RFS

P-value

HR

95% CI for HR

0.12

1.00 1.16

o0.01

1.00 0.67

NRM

REL

P-value

HR

95% CI for HR

P-value

HR

95% CI for HR

P-value

HR

95% CI for HR

0.96–1.40

0.41

1.00 1.08

0.90–1.28

o0.01

1.00 1.40

1.12–1.75

0.02

1.00 0.70

0.53–0.94

0.53–0.85

o0.01

1.00 0.72

0.58–0.90

o0.01

1.00 0.65

0.47–0.89

0.19

1.00 0.81

0.59–1.11

Male subjects with a female donor (F-M) had significantly lower relapse rate than male subjects with a male donor (M-M), while females had no significant difference in relapse rate due to the donor gender. OS ¼ overall survival; RFS ¼ relapse-free survival; NRM ¼ nonrelapse mortality; REL ¼ relapse rate; HR ¼ Hazard ratio; CI ¼ confidence interval. Adjustment factors as seen in Statistical Methods.

For each variable with missing values, it was verified that the outcomes of the ‘missing’ group were statistically equivalent to the outcomes of the ‘known’ group (all the other levels considered as a unique one), indicating that there was no hidden pattern of risk factors in the missing group. As a consequence, including the latter as one of the categories of the variable allowed the estimation of the adjusted effect of gender combination without problems in the interpretation and with the advantage of keeping all cases in the analysis. For each outcome, the candidate adjustment variables were: age, previous autologous transplantation, stage at diagnosis, time interval between diagnosis and transplant, response to chemotherapy, status at transplantation, source of stem cells, beta2-microglobulin levels, type of myeloma, TBI, T-cell depletion, reduced-intensity conditioning regimen (RIC), and calendar year (as a linear effect). The number of treatment lines before transplantation was not included as a candidate variable, firstly because it correlated significantly with the time interval from diagnosis to transplant and secondly since the information was lacking in 39% of the patients. The selection of the adjustment variables was carried out applying a backward procedure (except for the calendar year which was included regardless of the significance of its effect to represent possible unmeasured sources of heterogeneity correlated to it). No interaction between donor–recipient gender combination and any other variable was significant. For each model, the proportionality of the hazards was assessed, and the models for RFS and REL were stratified with respect to response status at transplant to avoid violation of this assumption of the Cox model. In order to focus on the outcomes of patients transplanted before/after 1994, we considered a follow-up truncated at 2 years (a time period equal to the shorter median follow-up among the four donor–recipient groups for patients transplanted from 1994 onwards). Owing to the restriction of the follow-up period, the analysis of RELs would not have been informative, and therefore it was not performed. We also analyzed the occurrence of acute graft versus host disease (aGVHD) with respect to the calendar period. The donor gender impact on REL, NRM, and OS in patients that had received the transplant from 1994 onwards was also analyzed separately, without restriction of follow-up. Bone Marrow Transplantation

The statistical analyses were performed using SPSS version 10.0.7 (2000), except the estimation of the cumulative incidence and the Gray test which were performed using R 1.6.2 (2003) and the software ‘cmprsk’ by T Gray, version 2 1-2 (2000).

Results Engraftment Engraftment was evaluable for 97% of the population. The engraftment frequency was similar in all groups, that is, 92.4, 91.3, 90.3, and 93.0%, respectively, in M-M, F-M, M-F, and F-F (comparison among male patients: P ¼ 0.8; comparison among female patients: P ¼ 0.5). Although there was a small tendency for a higher rejection rate in M-F than in F-F in evaluable patients (6.8 vs 5.0%), this difference was not statistically significant. Also, the time to engraftment did not differ significantly between the groups (P ¼ 0.6). The median time to neutrophils 0.5  106/l was 16 days and the median time to platelets 50  106/l was 31 days.

Response to transplantation There was no significant difference between the groups in the estimated minimum CR rates at 2 years, which were 40, 42, 38, and 42% for F-M, M-M, M-F, and F-F, respectively. The P-value for the global effect of donor gender was 0.92 (F-M vs M-M: P ¼ 0.63; M-F vs F-F: P ¼ 0.80). As illustrated in the statistical section, the method applied for the estimation of CR rate produces only minimum estimates, which therefore were lower than CR rates reported previously.26

Survival and NRM Male patients had a poor survival irrespective of whether the donor was male or female (Figure 1). However, there was no significant difference in OS between F-M and M-M, neither in the univariate (P ¼ 0.20) nor the multivariate analyses (P ¼ 0.12) (Table 2). Still, F-M had a much higher NRM than M-M (Po0.01) (Figure 2, Table 2).

Donor gender in transplantation of myeloma G Gahrton et al

613 1.0

1.0

0.8

0.8

0.6

M

M

F M

M F

F

F

12

24

36

0.6

0.4

0.4

0.2

M

M

F M

M F

F

F

0.0 0

12

24

0.2

36

48

60

72

84

96

108

120

23

18

10

0.0

Time (months)

0

M

M

476

234

162

106

77

53

38

27

48 60 72 Time (months)

84

96

108 120

F

M

334

141

91

54

33

22

15

10

5

1

0

M

M 476

196

116

70

54

40

28

19

14

12

7

M

F

258

122

76

49

30

23

17

12

10

9

8

F

M 334

122

71

40

21

15

10

8

5

1

0

F

F

244

136

100

75

58

40

31

28

18

16

11

M

F

258

102

63

37

23

14

10

9

6

5

4

F

F

244

114

82

58

45

27

19

17

10

7

5

Population at risk

Population at risk

Figure 1 Kaplan–Meier curves for OS. Better OS in female subjects with a female donor (F-F) than in female subjects with a male donor (M-F) (Po0.01) while there is no significant difference between male subjects with a female (F-M) or a male (M-M) donor.

The survival for F-F was significantly better than for M-F (Po0.01). The combination was superior to all other donor–patient combinations (Figure 1). The reason for the difference was a lower NRM in F-F as compared to M-F (Po0.01) (Table 2, Figure 2). Owing to the previously reported improvement in survival in patients transplanted from 1994,26 an analysis of the donor influence on patients transplanted before and from 1994 was made (Figure 3a and b, Table 3). For comparability this required restriction of the follow-up period to the first 2 years after transplant (see Statistical analysis). Before 1994, the OS was significantly poorer in F-M than in M-M but similar in both groups from 1994 and onwards (Figure 3a). NRM was significantly reduced during the later time period, but comparatively more in F-M than in M-M, partly explaining the similar OS in F-M and M-M from 1994 (both combinations significantly improved as compared to the time period before 1994 as previously shown26). In female patients NRM was lower in F-F as compared to M-F in patients transplanted from 1994 but on an improved level as compared to transplants before 1994. The difference in OS between F-F and M-F was also similar to the difference before 1994 but on an improved level, F-F still being superior to M-F (Figure 3b).

Figure 2 Cumulative incidence of NRM. Higher NRM in male patients with a female donor (F-M) as compared to male patients with a male donor (M-M) (Po0.01) and in female subjects with a male donor (M-F) as compared to female subjects with a female donor (F-F) (Po0.01).

F-F (9.3%) (P ¼ 0.096). This tendency was even more pronounced in transplants performed since 1994 (M-F 16.1% vs F-F 9.2%; P ¼ 0.06). Although the incidence of severe aGVHD in female subjects tended to be higher with a male donor, GVHD was not reported to be the primary cause of death more frequently than in female subjects with a female donor (Table 5). In male patients, there was no significant difference (P ¼ 0.447) in severe aGVHD between those who received the graft from a male (14.4%) as compared to a female donor (16.4%) nor was aGVHD reported to be the primary cause of death more frequently in F-M (22.8%) than in M-M (16.7%) (P ¼ 0.152). The same lack of difference in aGVHD between the two groups F-M and M-M was seen in transplants performed after 1993 (data not shown). The frequency of chronic GVHD (cGVHD) is seen in Table 6. While for female patients there was no association between donor gender and the incidence or severity of cGVHD (P ¼ 0.986), male patients had significantly (P ¼ 0.006) more cGVHD with a female donor (limited 23% and extensive 33%) than with a male donor (limited 17%, extensive 22%).

REL and RFS GVHD As seen in Table 4, severe aGVHD grades III–IV tended to be more frequent in M-F (14.3%) as compared to

The REL is seen in Figure 4. The REL was significantly lower in F-M as compared to M-M (univariate analysis: P ¼ 0.002; multivariate: P ¼ 0.02) (Table 2). This corroboBone Marrow Transplantation

Donor gender in transplantation of myeloma G Gahrton et al

614

a

b

1.0

1.0

0.8

0.8

0.6

0.6

0.4

0.4 Female patients

Male patients Male donor, =1994

0.0

57

57

18

20

22

0

24

2

4

6

8

10

12

14

16

18

20

22

24 26

Time (months) 55

51

50

50

F > 94 72 47 39 31 28 26 25 23 23 22 22 22 22 M ≥ 94 357 300 253 232 208 192 177 160 147 136 128 117 112 F ≥ 94 262 219 183 155 140 128 116 111 104 93 83 76 69

60

46

38

34

33

31

31

31

31

30

29

29

69

69

62

55

53

48

47

46

43

42

40

39

38

111 99

91

90

69

63

57

50

F ≥ 94 139 139 118 109 103 95

89

86

78 81

73

70

67

62

M < 94 F > 94

M ≥ 94 168 168 140 123

Population at risk

Population at risk

Figure 3

(a) Male patients: for the OS in the first 2 years, the female donor is associated with a higher mortality only for patients transplanted before 1994 (95% CI: 1.03–2.17) and not for patients transplanted from 1994 (95% CI for HR: 0.8–1.4). (b) Female patients: for the OS in the first 2 years the female donor is superior to the male donor for patients transplanted after 1993 (95% CI for HR: 0.5–0.9).

Table 3

Effect of gender combination (donor-recipient) before and since 1994 (multivariate analysis) OS at 2 years before 1994

OS at 2 years since 1994

NRM at 2 years before 1994

NRM at 2 years since 1994

P-value

HR

95% CI for HR

P-value

HR

95% CI for HR

P-value

HR

95% CI for HR

P-value

HR

95% CI for HR

M-M F-M

0.03

1.00 1.50

1.03–2.17

0.65

1.00 1.06

0.83–1.36

0.01

1.00 1.73

1.16–2.59

0.12

1.00 1.25

0.94–1.66

M-F F-F

0.18

1.00 0.73

0.46–1.16

0.02

1.00 0.68

0.49–0.94

0.17

1.00 0.68

0.39–1.18

0.02

1.00 0.62

0.42–0.94

Overall survival (OS) was poorer and nonrelapse mortality (NRM) higher in F-M vs (M-M) before 1994, while there was no significant difference between these parameter from 1994 onwards. The outcomes are truncated at 2 years for comparability. Adjustment factors as seen in Statistical methods.

rates the observation that relapse as a cause of death was more frequent in M-M (47.1%) as compared to F-M (29.0%). The lower REL compensated for the higher NRM in F-M, thus OS and RFS were similar in M-M and F-M (see above) (Figures 1 and 5 and Table 2). However as seen in Figure 3a, in patients transplanted before 1994, there was a significant difference in OS to the disadvantage of F-M due to the higher NRM as compared to M-M, in contrast to the finding in F-M during the later time period when NRM and OS were similar in F-M and MM (Table 3). Although the REL for these two periods could not be compared beyond 2 years from transplant (see statistical methods), it was noted that it was 28% in M-M and 20% in F-M at this time. Also, a separate comparison of the relapse risk in F-M and M-M in Bone Marrow Transplantation

patients transplanted from 1994 and onwards, disregarding the comparison between time periods, showed that F-M had a lower relapse risk (HR: 0.64, 95% CI: 0.46–0.90, P ¼ 0.009). In female subjects, there was no significant difference in REL (univariate analysis: P ¼ 0.824; multivariate analysis: P ¼ 0.19), and no difference in relapse as a cause of death due to the donor gender (47.9% of M-F and 51.5% of F-F, P ¼ 0.549) . This was also true for the REL in the separate analysis in transplants performed after 1993 (HR: 0.73, 95% CI: 0.50–1.08, P ¼ 0.119). Thus, the REL did not seem to have an impact on the advantageous OS (Figure 1) and RFS (Figure 5) seen in F-F as compared to M-F, either in transplants performed before or from 1994 (data not shown). These differences were due to differences in NRM.

Donor gender in transplantation of myeloma G Gahrton et al

615

Discussion This study clearly demonstrates that transplantation of a male patient with a graft from a female donor involves a graft vs myeloma effect. This donor gender-specific effect has not been seen in previous studies of myeloma transplants, probably because of lack of detectable advantageous effect on OS, masked by increased transplant-related mortality, and inadequate follow-up. Before 1994, there was even a highly significantly poorer OS in F-M than in M-M due to this higher NRM hiding the impact of a reduced REL. However with the decrease in this mortality after 1993, unmasking the reduced REL, the OS was improved and became similar in M-M and F-M. The lower REL in F-M corroborates with more cGVHD in these patients compared to M-M. An association between graft vs myeloma and GVHD has previously been reported.13,30–32 Recent studies in chronic myelocytic leukemia and acute leukemia have shown a similar gender-dependent effect, that is, lower REL and higher transplant-related mortality in female to male as compared to male to male transplants and this was assumed to be due to the presence of female donor T cells that are specific for recipient minor

Table 4

Acute GVHD

histocompatibility antigens (H-Y) encoded by Y chromosome genes.24,33 H-Y antigens have been shown to be present in a variety of male cells and anti H-Y peptidespecific cytotoxic T cells as well as H-Y-directed antibodies have been identified in male patients who received marrow from female donors.25 Women may be sensitized to men during pregnancy34,35 and therefore some female patients may have cytotoxic T cells directed against male antigens before the donation. An association between GVHD and presence of cytotoxic T cells was found in a woman who rejected bone marrow from a male donor.20–23 Thus it seems reasonable to conclude that H-Y antigens are responsible both for eliciting a stronger graft vs myeloma effect in male patients with a female donor than with a male donor, and the concomitant higher GVHD related mortality in this donor–patient combination. By decreasing overall transplant-related mortality with improved methods of preventing transplant-related complications during the years from 1994, it has been possible to unmask the lower REL, translating into a significantly better survival in this group as compared to the earlier time period. The present study confirms previous observation in multiple myeloma that the combination female recipient– female donor has the best outcome. No advantageous impact of gender disparity could be seen on REL, RFS, or

Chronic GVHD

Table 6 Grade of aGVHD

Total

Grade of cGVHD

Grades 0–2

Grades 3–4

Male patients (P ¼ 0.447) Male donor (M-M) Female donor (F-M) Total

369 (86%) 249 (84%) 618 (85%)

62 (14%) 49 (16%) 111 (15%)

431 298 729

Male patients (P ¼ 0.006) Male donor (M-M) Female donor (F-M) Total

Female patients (P ¼ 0.096) Male donor (M-F) Female donor (F-F) Total

204 (86%) 205 (91%) 409 (88%)

34 (14%) 21 (9%) 55 (12%)

238 226 464

Grade of acute graft-versus-host disease (aGVHD) with respect to donor gender in male and female patients (with P-values for comparison). There was no significant difference between M-M vs F-M, but a tendency for more severe aGVHD in M-F than in F-F (P ¼ 0.096). Comparison of other subgroups of aGVHD did not show significant differences related to the donor gender, that is, grades 0–I and II–IV were not significantly different in male subjects (P ¼ 0.612) or female subjects (P ¼ 0.171).

Table 5

No

Limited

Extensive

122 (61%) 53 (44%) 175 (55%)

34 (17%) 27 (23%) 61 (19%)

44 (22%) 39 (33%) 83 (26%)

200 119 319

Female patients (P ¼ 0.986) Male donor (M-F) 44 (46%) Female donor (F-F) 46 (45%) Total 90 (46%)

32 (33%) 33 (33%) 65 (33%)

20 (21%) 22 (22%) 42 (21%)

96 101 197

Maximum grade of chronic graft-versus-host disease (cGVHD) with respect to donor gender in male and female patients (with P-values for comparison). There was significantly less limited and extensive (cGVHD) in M-M compared to F-M (P ¼ 0.006), but no significant difference between M-F and F-F.

Causes of death

Total number of deaths

Infections GVHD Relapse Other

Total

M-M

F-M

M-F

F-F

Total

221

145

119

97

582

Number

%

Number

%

Number

%

Number

%

Number

%

41 37 104 39

18.6 16.7 47.1 17.6

35 33 42 35

24.1 22.8 29.0 24.1

22 14 57 26

18.5 11.8 47.9 21.8

17 11 50 19

17.5 11.3 51.5 19.6

115 95 253 119

19.8 16.3 43.5 20.4

The global chi-squared test has a P-value of 0.017 meaning that the causes of death are different in the four groups. Among male patients, F-M has significantly less deaths due to relapse (29.0%) than M-M (47.1%) (P ¼ 0.007). Among female patients, there was no significant difference between M-F and F-F (P ¼ 0.968). Bone Marrow Transplantation

Donor gender in transplantation of myeloma G Gahrton et al

616 1.0

1.0

M F M F

0.8

M M F F

0.8

0.6

0.6

0.4

0.4 M F M F

0.2

0.2 0.0

0.0

0

0

12

24

36

48

60

72

84

96 108 120

Time (months)

12

108 120

70

54

40

28

19

14

12

7

40 37

21 23

15 14

10 10

8 9

5 6

1 5

0 4

82

58

45

27

19

17

10

7

5

54

40

28

19

14

12

7

M

F

71

40

21

15

10

8

5

1

0

F

F 244 114

M

M 334 122 F 258 102

63

37

23

14

10

9

6

5

4

F

F 244 114

82 58

45

27

19

17

10

7

5

Bone Marrow Transplantation

96

71 63

70

OS in female subjects. Thus, there seem to be no minor histocompatibility antigens in female subjects eliciting a clinically detectable antimyeloma response from male T cells. However, it is interesting to note that 6.6% of the female subjects rejected male grafts, but only 4.9% of the female grafts. Although not significant, this tendency for more rejections of male grafts in female subjects corroborates with an importance of male H-Y antigens, perhaps inducing an immune response in some female subjects. However, there is no obvious explanation for the significantly lower NRM in female to female transplants as compared to male to female ones. Female to female transplants have the best OS of all four combinations. Multiple myeloma appears to be a good model to explore the possible advantageous effect of H-Y as well as other minor histocompatibility antigens. Donor lymphocyte transfusions in relapsed patients may induce up to 30% CRs.36 Improving the specificity of donor lymphocytes by sorting or utilizing differences in minor histocompatibility antigens in other ways may prove to be beneficial. With more effective prevention and treatment of transplantrelated complications, the effect of a donor gender mismatch in male patients could be further explored. The present study has only utilized sibling donor transplants in order to prove the principle. It may well be possible that similar effects could be explored in unrelated transplants. From the practical point of view, it may be more realistic to

48 60 72 84 Time (months)

F

M 476 196 116

Cumulative incidence of relapse. Significantly lower incidence of relapse in male subjects with a female donor (F-M) than in male subjects with a male donor (M-M) (P ¼ 0.02), but no difference in female subjects with a male (M-F) or a female (F-F) donor (P ¼ 0.19).

36

M 476 196 116

M

Figure 4

24

M

M 334 122 F 258 102

Population at risk

M M F F

Population at risk

Figure 5 Kaplan–Meier curves for RFS. Significantly better RFS in female subjects with a female donor (F-F) than in female subjects with a male donor (M-F) (Po0.01), but no difference between male subjects with a female (F-M) or a male donor (M-M) (P ¼ 0.41).

select for gender-specific minor histocompatibility antigens in the unrelated donor pool, now containing more than 8 million HLA-typed donors.37 Within a family it is rare to find two major histocompatibility antigen compatible donors with opposing gender. With further improvement in NRM it may well be advantageous to choose a female instead of a male donor for a male patients, in contrast to what has been thought until now. Presently, there seems to be no reason to preferentially choose a male donor.

Acknowledgements This study was supported by grants from the Swedish Cancer Fund and the Cancer Society in Stockholm. This work is based on data reported to the EBMT registry by EBMT centers , which is greatly acknowledged.

References 1 Alexanian R, Haut A, Khan AU et al. Treatment for multiple myeloma. Combination chemotherapy with different melphalan dose regimens. JAMA 1969; 208: 1680–1685. 2 Sporn JR, McIntyre OR. Chemotherapy of previously untreated multiple myeloma patients: an analysis of recent treatment results. Semin Oncol 1986; 13: 318–325. 3 Myeloma Trialists’ Collaborative Group. Combination chemotherapy versus melphalan plus prednisone as treatment for multiple myeloma: an overview of 6,633 patients from 27 randomized trials. J Clin Oncol 1998; 16: 3832–3842.

Donor gender in transplantation of myeloma G Gahrton et al

617 4 Blade J, San Miguel JF, Fontanillas M et al. Increased conventional chemotherapy does not improve survival in multiple myeloma: long-term results of two PETHEMA trials including 914 patients. Hematol J 2001; 2: 272–278. 5 Barlogie B, Alexanian R, Dicke KA et al. High-dose chemoradiotherapy and autologous bone marrow transplantation for resistant multiple myeloma. Blood 1987; 70: 869–872. 6 Child JA, Morgan GJ, Davies FE et al. High-dose chemotherapy with hematopoietic stem-cell rescue for multiple myeloma. N Engl J Med 2003; 348: 1875–1883. 7 Attal M, Harousseau JL, Stoppa AM et al. A prospective, randomized trial of autologous bone marrow transplantation and chemotherapy in multiple myeloma. Intergroupe Francais du Myelome. N Engl J Med 1996; 335: 91–97. 8 Gahrton G, Ringden O, Lonnqvist B et al. Bone marrow transplantation in three patients with multiple myeloma. Acta Med Scand 1986; 219: 523–527. 9 Tura S, Cavo M, Baccarani M et al. Bone marrow transplantation in multiple myeloma. Scand J Haematol 1986; 36: 176– 179. 10 Gahrton G, Tura S, Ljungman P et al. Allogeneic bone marrow transplantation in multiple myeloma. European Group for Bone Marrow Transplantation. N Engl J Med 1991; 325: 1267–1273. 11 Bjorkstrand BB, Ljungman P, Svensson H et al. Allogeneic bone marrow transplantation versus autologous stem cell transplantation in multiple myeloma: a retrospective casematched study from the European Group for Blood and Marrow Transplantation. Blood 1996; 88: 4711–4718. 12 Gahrton G, Tura S, Ljungman P et al. Prognostic factors in allogeneic bone marrow transplantation for multiple myeloma. J Clin Oncol 1995; 13: 1312–1322. 13 Corradini P, Cavo M, Lokhorst H et al. Molecular remission after myeloablative allogeneic stem cell transplantation predicts a better relapse-free survival in patients with multiple myeloma. Blood 2003; 102: 1927–1929. 14 Bensinger WI, Buckner CD, Anasetti C et al. Allogeneic marrow transplantation for multiple myeloma: an analysis of risk factors on outcome. Blood 1996; 88: 2787–2793. 15 Bensinger WI, Maloney D, Storb R. Allogeneic hematopoietic cell transplantation for multiple myeloma. Semin Hematol 2001; 38: 243–249. 16 Sasazuki T, Juji T, Morishima Y et al. Effect of matching of class I HLA alleles on clinical outcome after transplantation of hematopoietic stem cells from an unrelated donor. Japan Marrow Donor Program. N Engl J Med 1998; 339: 1177–1185. 17 Gahrton G, Tura S, Ljungman P et al. An update of prognostic factors for allogeneic bone marrow transplantation in multiple myeloma using matched sibling donors. European Group for Blood and Marrow Transplantation. Stem Cells 1995; 13 (Suppl 2): 122–125. 18 Zwaan FE, Hermans J, Barrett AJ et al. Bone marrow transplantation for acute nonlymphoblastic leukaemia: a survey of the European Group for Bone Marrow Transplantation (E.G.B.M.T.). Br J Haematol 1984; 56: 645–653. 19 Gratwohl A, Hermans J, Goldman JM et al. Risk assessment for patients with chronic myeloid leukaemia before allogeneic blood or marrow transplantation. Chronic Leukemia Working Party of the European Group for Blood and Marrow Transplantation. Lancet 1998; 352: 1087–1092. 20 Vogt MH, de Paus RA, Voogt PJ et al. DFFRY codes for a new human male-specific minor transplantation antigen involved in bone marrow graft rejection. Blood 2000; 95: 1100–1105.

21 Vogt MH, Goulmy E, Kloosterboer FM et al. UTY gene codes for an HLA-B60-restricted human male-specific minor histocompatibility antigen involved in stem cell graft rejection: characterization of the critical polymorphic amino acid residues for T-cell recognition. Blood 2000; 96: 3126–3132. 22 Vogt MH, van den Muijsenberg JW, Goulmy E et al. The DBY gene codes for an HLA-DQ5-restricted human malespecific minor histocompatibility antigen involved in graftversus-host disease. Blood 2002; 99: 3027–3032. 23 Spierings E, Vermeulen CJ, Vogt MH et al. Identification of HLA class II-restricted H-Y-specific T-helper epitope evoking CD4+ T-helper cells in H-Y-mismatched transplantation. Lancet 2003; 362: 610–615. 24 Randolph SS, Gooley TA, Warren EH et al. Female donors contribute to a selective graft-versus-leukemia effect in male recipients of HLA-matched, related hematopoietic stem cell transplants. Blood 2004; 103: 347–352. 25 Miklos DB, Kim HT, Zorn E et al. Antibody response to DBY minor histocompatibility antigen is induced after allogeneic stem cell transplantation and in healthy female donors. Blood 2004; 103: 353–359. 26 Gahrton G, Svensson H, Cavo M et al. Progress in allogenic bone marrow and peripheral blood stem cell transplantation for multiple myeloma: a comparison between transplants performed 1983–93 and 1994–8 at European Group for Blood and Marrow Transplantation centres. Br J Haematol 2001; 113: 209–216. 27 Gray R. A class of k-sample tests for comparing the cumulative incidence of competing risk. Ann Statist 1988; 16: 1141–1154. 28 Prentice RL, Kalbfleisch JD, Peterson Jr AV et al. The analysis of failure times in the presence of competing risks. Biometrics 1978; 34: 541–554. 29 Gooley TA, Leisenring W, Crowley J et al. Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat Med 1999; 18: 695–706. 30 Aschan J, Lonnqvist B, Ringden O et al. Graft-versusmyeloma effect. Lancet 1996; 348: 346. 31 Lokhorst HM, Schattenberg A, Cornelissen JJ et al. Donor lymphocyte infusions for relapsed multiple myeloma after allogeneic stem-cell transplantation: predictive factors for response and long-term outcome. J Clin Oncol 2000; 18: 3031–3037. 32 Tricot G, Vesole DH, Jagannath S et al. Graft-versus-myeloma effect: proof of principle. Blood 1996; 87: 1196–1198. 33 Gratwohl A HJ, Niederwieser D, van Biezen A et al. Female donors influence transplant-related mortality and relapse incidence in male recipients of sibling blood and marrow transplants. Hematol J 2001; 2: 363–370. 34 James E, Chai JG, Dewchand H et al. Multiparity induces priming to male-specific minor histocompatibility antigen, HY, in mice and humans. Blood 2003; 102: 388–393. 35 Verdijk RM, Kloosterman A, Pool J et al. Pregnancy induces minor histocompatibility antigen-specific cytotoxic T cells: implications for stem cell transplantation and immunotherapy. Blood 2004; 103: 1961–1964. 36 Lokhorst HM, Schattenberg A, Cornelissen JJ et al. Donor leukocyte infusions are effective in relapsed multiple myeloma after allogeneic bone marrow transplantation. Blood 1997; 90: 4206–4211. 37 WMDA. Donor Registries Annual Report 2002, Vol 1, 6th edn. WMDA: Leiden, The Netherlands, 2003.

Bone Marrow Transplantation