Problems and possible solutions in finding an unrelated bone ... - Nature

Bone Marrow Transplantation, (1997) 20, 1011–1017  1997 Stockton Press All rights reserved 0268–3369/97 $12.00

Problems and possible solutions in finding an unrelated bone marrow donor. Results of consecutive searches for 240 Dutch patients M Oudshoorn 1,2, JJ Cornelissen3 , WE Fibbe4, ER de Graeff-Meeder5, JLWT Lie2, GMTh Schreuder1, K Sintnicolaas6, R Willemze4, JMJJ Vossen7 and JJ van Rood1,2 1

Department of Immunohaematology and Blood Bank, 2Europdonor Foundation, 4Department of Haematology, 7Department of Paediatrics, Leiden University Medical Centre, Leiden; 3Department of Haematology, Dr Daniel den Hoed Cancer Centre, Rotterdam; 5Department of Immunology/Haematology, University Children’s Hospital, ‘Het Wilhelmina Kinderziekenhuis’, Utrecht; and 6Red Cross Blood Bank Rotterdam, Rotterdam, The Netherlands

Summary: To evaluate the efficiency of our protocol for finding an HLA matched unrelated bone marrow donor, search results obtained between 1990 and 1995 for 240 Dutch patients were analyzed. The percentage of patients for whom, according to information given by the registries, a fully split-HLA antigen matched donor is available, increased from 24% in 1990 to over 70% in 1995. As a result the percentage of patients transplanted rose from about 24% in 1990–1991 to 44% in 1994–1995. The median time between the start of the search and transplantation was about 6 months. The systematic use of Bone Marrow Donors Worldwide (BMDW) which comprises the HLA groups of all volunteer bone marrow donors in Europe, Israel, South Africa, North America, Canada, India, Australia and New Zealand has been essential in this context. While searching for a suitable donor several problems were encountered such as unavailability of donors (12%) and discordant typing results (8%; range ,1% to .25%). Thus it is advisable to select several donors for a patient. For 86% of patients with at least one HLA identical donor on the serological level for HLA-A,-B,-DR,-DQ, an HLADRB1/3/4/5, and -DQB1 identical donor could be identified. As expected, patients with two frequent haplotypes in strong linkage disequilibrium had the best chance of obtaining an HLA matched donor. Unexpectedly, patients with only one such haplotype had an almost similar chance. It could be calculated that HLA-DR typing of HLA-A,-B identical donors was rarely costeffective after 1992. Only 12 of the 75 transplanted patients (16%) typeable at DNA level for class II, turned out to be completely matched for HLA-A,-B,-C,DRB1/3/4/5,-DQB1,-DPB1 and had a negative MLC test. In the group of patients transplanted with a fully matched donor and for whom a CTLp test was performed, only 7% (4/54) of the tests were negative. Search results for patients of non-European origin were dismal, with only four of 26 patients referred being transCorrespondence: M Oudshoorn, Europdonor Foundation, Bldg 1, E3-Q, Leiden University Medical Centre, POB 9600, 2300 RC Leiden, The Netherlands Received 24 April 1997; accepted 14 July 1997

planted. In summary, of the 240 patients for whom the Europdonor office searched for a donor, about onethird were transplanted, one-third had a potential donor but did not reach transplantation, while for the remaining one-third of patients no suitable donor could be found. Keywords: unrelated donors; BMT; donor search

Only one out of three patients in need of a bone marrow transplant has a suitable sibling donor. Allogeneic bone marrow transplantation with marrow from unrelated but HLA matched donors has proven to be a suitable alternative for such patients.1,2 Small and large registries have been formed to facilitate the search for an unrelated donor.3–7 Over half of potential donors are coordinated by the National Marrow Donor Program (NMDP) and approximately 40% are included in the European Donor Secretariat (EDS) or European Marrow Donor Information System (EMDIS). According to the IMUST study8 a suitable donor can be found for 10% of patients only, if the search is restricted to a large registry such as the British Bone Marrow and Platelet Donor Panel (BBMPDP) with 40 000 donors at the time of analysis. This figure is much higher if other registries are contacted as well.9–12 Selection criteria by which a suitable donor is identified, are still in a state of flux as a result of the introduction of new HLA typing technologies, extension of the registries and the introduction of new bone marrow transplant protocols. In this report we describe the results of search requests for 240 Dutch patients reported to Europdonor in an attempt to document successes and failures of the search process, and if possible their causes.

Patients and methods The analysis concerned the donor searches for 240 consecutive Dutch patients registered by the Europdonor Foundation from 1990 till 1995. Patients originated from the following transplant centres: the University Hospital Leiden (adults n = 50 and children n = 105), Dr Daniel den Hoed Cancer Centre Rotterdam (n = 64) and the University Chil-

Finding an unrelated bone marrow donor M Oudshoorn et al

1012

Table 1

Patients’ characteristics

No. of patients Sex (male/female) Age (years) median range under 16 years/over 16 years

240 144/96 14.4 0.1–52.1 131/109

Diagnosis (No. of patients) Chronic myelogeneous leukaemia Acute myelogeneous leukaemia Acute lymphocytic leukaemia Other malignant diseases Severe aplastic anaemia Fanconi anaemia Myelodysplastic syndromes Severe combined immune deficiency Other non-malignant diseases

51 41 39 6 31 10 25 10 27

dren’s Hospital, ‘het Wilhelmina Kinderziekenhuis’ in Utrecht (n = 21). All patients were in need of an allogeneic bone marrow transplant but had no HLA identical sibling. Patients’ characteristics shown in Table 1 did not differ from other studies. Search protocol The outline of the process for selecting a bone marrow donor for patients in need of an allogeneic bone marrow donor but lacking an HLA (genotypically) identical sibling donor is given in Figure 1.

After a patient has been reported to Europdonor, the HLA data of his/her siblings and parents are requested in order to establish the patient’s haplotypes. The patient is retyped in the national reference laboratory for histocompatibility. If the patient has a haplotype with a high frequency the chance of finding a phenotypically identical family member (eg cousins, aunts, uncles etc) is calculated.13 If the chance is around 5–10% or higher a search for an extended family donor is initiated. A search for an extended family donor is also initiated if there is a consanguinous marriage in the patient’s family. At the same time, Bone Marrow Donors Worldwide (BMDW) which contains the HLA phenotypes and number of donors per phenotype from bone marrow donor registries all over the world is consulted.14 The 2nd edition (March 1990) to 25th edition (December 1995) is screened for donors with suitable phenotypes. In consultation with the transplant centres a selection of (a) donor(s) is made. The first choice of unrelated donors are HLA-A,-B, -DRB1 typed donors, second HLA-A,-B,-DR typed donors identical at the split level to the patient, and third, donors who are as yet not typed at the level of the split antigens but are identical for the broad HLA-A,-B,-DR antigens. If only HLA-AB identical donors are reported in BMDW, the chance that HLA-DR typing will identify an HLA-A,-B,DR identical donor is calculated.15 If the chance is higher than 1%, which implies that if 10 donors are HLA-DR typed there is a chance of almost 10% that one donor is HLA-A,-B and -DR identical, HLA-DR typing is initiated.

I. Patient reported to ED II. HLA-retype patient III. Establish haplotypes of patient IV. Consider extended family search

Va. Extended family search

Often done

Vb. Search for unrelated donor

simultaneously

Donor*

Donor*

No donor

Vb.

VI. Only AB id. donor(s)

HLA-DR typing

VII. No donor

Search 1 antigen mismatch donor

*Histocompatibility testing HLA retyping (serol) DNA typing (HLA-A2,C,DRB1/3/4/5, DQB1,DPB1) MLC test CTLp test

Figure 1

Outline of the process for selecting a bone marrow donor.

Donor*

No donor VII

Donor*

No donor

Finding an unrelated bone marrow donor M Oudshoorn et al

If HLA identical donors are not available, a search program is used which identifies donors who have one antigen mismatched at the split level or a cross-reacting antigen mismatch (eg HLA-A24 instead of -A23, HLA-A2 instead of -A28, respectively). At the same time, the chance of finding a one locus antigen mismatched family donor is calculated.13 All donors requested for a patient are retyped using serological methods. From 1992 onwards, patient and donors have also been subtyped at DNA level for HLADRB1/3/4/5 and increasingly for HLA-DQB1, -DPB1, -A2 and -C. The mixed lymphocyte culture test is performed routinely and the cytotoxic T lymphocyte precursor (CTLp) test performed if possible. HLA serology All HLA-A,-B and -C typings of patients and retypings of donors were done using the standard NIH lymphocytotoxicity assay and the HLA-DR and -DQ was performed with the two colour fluorescence assay16 using allo-antisera.

requested. Twelve percent of donors requested for testing were unavailable due to medical or personal reasons (Table 3). Following retyping, discordant results of 8% were observed (Table 3). This number did not change during the period studied. It did, however, vary between the registries from less than 1% to over 25% (mean = 8%; median = 11%). Donors from registries with more than 50 000 donors had on average more discordant results (mean = 10%; median = 12%) than registries with less than 50 000 donors (mean = 3%; median 8%). In 37 out of 73 (51%), the discordant results involved homozygosity, ie donor centre typed the donor as homozygous while the donor was heterozygous and visa versa, ie the donor was typed as heterozygous but was in fact homozygous for that locus. For 8% of patients the only donor thought to be present, turned out to be either not available (4%) or mistyped (4%).

Percentage of patients for whom an HLA-A,-B, -DRB1/3/4/5,-DQB1 identical donor could be found

DNA typing HLA-DRB1 typing was performed using biotinylated sequence-specific oligonucleotides from 1992 to 199317 and from 1994 to 1995 using the PCR-SSP technique.18 The HLA-DPB1 typing was done with similar methods using three primer sets, one generic set and the other two separating the DPB1 specificities into two groups. Twenty-one DPB1 sequence-specific oligonucleotides were used to detect 36 different DPB1 specificities. HLA-C typing was performed according to the method of Bunce et al.19 Whenever possible, patient/donor pairs not yet typed at the DNA level, were typed retrospectively. Mixed lymphocyte culture (MLC) test The mixed lymphocyte cultures were performed according to Keuning et al.20 Tests were defined as negative when the relative response value was equal to or less than 10%. Cytotoxic T lymphocyte precursor (CTLp) test The CTLp test was performed as described by Zhang et al21 with minor modifications. In each test, donor cells were used as responder cells and patient cells as stimulator and target cells. A negative CTLp test was defined as less than 2 CTLps per 106 peripheral blood lymphocytes, a doubtful positive result was defined as >2 and ,8 and a positive result as >8 CTLps per 106 peripheral blood lymphocytes. Results Number of potentially available donors Table 2 illustrates the rise of potentially HLA-A,-B,-DR identical donors (at the split level) available for patients in the years 1990–1995 in either the Dutch Europdonor registry or in BMDW. From these potentially acceptable donors a mean number of 6.7 donor blood samples per patient were

Analysis is restricted to the years 1992–1995, because DNA typing was introduced as a routine technique in the middle of 1992. The data analysed consisted of 441 HLA-A,-B, -DR,-DQ serologically split matched donors tested for 97 patients. The mean number of donors tested per patient was 4.5; the median 5.0 and the range 1–17 donors. Seventyfive percent of all HLA-A,-B,-DR,-DQ split identical donors based on serology were also identical to the patient for HLA-DRB1/3/4/5 and 13% for HLA-DRB1/3/4/5, -DQB1 and -DPB1. For 86 of the 97 patients (89%) one (or more) HLA-,A,-B,-DRB1,-DQB1 identical donor(s) were identified. This figure was only reduced by 3% to 86% when identity for DRB3/4/5 was also included in the analysis. Of these, only 28% were also HLA-DPB1 identical.

Impact of linkage disequilibrium on the outcome of the search process To establish whether the linkage disequilibrium of the patients’ haplotypes would be helpful in predicting whether an HLA-A,-B,-DRB1/3/4/5,-DQB1 identical donor could be identified, the 97 patients described above were subdivided into those with two (FF), one (FN) or no haplotypes (NN) in strong linkage disequilibrium. A frequent haplotype in this analysis was defined as a haplotype belonging to the 10 haplotypes with strongest linkage disequilibrium in the Dutch population.22 The 11 patients with two haplotypes in strong linkage disequilibrium had, as expected, a very high success rate (100%) in locating an HLA-A,-B, -DRB1/3/4/5,-DQB1 identical donor, but also those patients (n = 45) who had only one such haplotype had a high success rate (91%). The chance of finding such a donor for patients (n = 41) with no haplotype in strong linkage disequilibrium was 78% (P = NS). The chance of finding an HLA-DPB1 identical donor was significantly increased in the FF group (63%; P = 0.04) compared to the combined FR and RR group (23%).

1013

Finding an unrelated bone marrow donor M Oudshoorn et al

1014

Table 2 The number of HLA-A,-B,-DR identical donors (typed at split level) available at time of initial search request in Europdonor and Bone Marrow Donors Worldwide (BMDW) for 240 Dutch patients Year in which search was started

Europdonor Foundation No. of donors registered typed at split level

1990 1991 1992 1993 1994 1995

n n n n n n

= = = = = =

42 28 56 28 36 50

No. of patients with a donor (% of total patients)

1430 1754 2942 5930 7977 10 250

n = 240 Total 1990–1995

BMDW

1 1 3 5 14 16

1430–10 250

(2%) (4%) (5%) (18%) (39%) (32%)

40 (17%)

No. of donors registered typed at split level

No. of patients with a donor (% of total patients)

26 181 48 932 64 638 179 389 332 764 515 651

10 11 30 16 27 36

26 181–515 651

(24%) (39%) (54%) (57%) (75%) (72%)

130 (54%)

n = No. of patients

Table 3

The number of HLA-A,-B,-DR typed donors requested, their availability and discordant HLA typing of the donors requested

Year in which search was started

No. of donor blood samples requesteda

1990 1991 1992 1993 1994 1995 1990–1995 a

No. of donors unavailableb

128 75 307 169 189 255 1123

9 7 30 19 26 44 135

(7%) (9%) (10%) (11%) (14%) (17%) (12%)

No. of donor samples retypedc 107 60 250 135 148 195 895

No. of donors with discordant HLA typing 8 4 25 11 9 16 73

(7%) (7%) (10%) (8%) (6%) (8%) (8%)

Not only HLA-A,-B,-DR split identical donors were requested but also broad and a few minor mismatched donors. Donor unavailable due to donor withdrawal for medical or personal reasons. Ninety-three donor samples received were not retyped.

b c

HLA-DR typing of HLA-A,-B identical donors HLA-DR typing was effective in the years 1990–1992 when 12 donors were identified for eight out of 30 patients after 215 HLA-DR typing procedures. In contrast, in the years 1993–1995, HLA-DR typing was requested sporadically and was always unsuccessful in identifying an HLAA,-B,-DR identical donor. Percentage of patients who did not reach transplantation and the reasons why Eighty-six (36%) out of the 240 patients were transplanted. Thirty-one patients (13%) were actively withdrawn from the search process, generally because of a reassessment of the indication for transplantation. Seventeen percent of patients died or were not transplantable due to advanced disease status. The percentage of patients who could not be transplanted because a donor was not found dropped from 50% in 1990–1991 to 21% in 1994–1995. The percentage of patients who could be transplanted rose from an average of 24% in 1990–1991 to 44% in 1994–1995. Donor–recipient disparities for HLA Although the a priori aim of the search was to find HLA split identical donors, a large percentage of patients actually

received an HLA mismatched graft. HLA typing at the DNA level was performed for some of the patient/donor combinations retrospectively. Most patient/donor pairs were found to be either mismatched for class I (A,B or C) and/or class II (DRB1/3/4/5,DQB1 or DPB1) or had a positive MLC or CTLp test. In the majority of the cases studied, the class I mismatches could be shown to be due to mismatches at the HLA-C locus (18/24). Twelve patients (16%) appeared to be transplanted with a well matched donor, ie identical for HLA-A,-B,-C,-DRB1/3/4/5,-DQB1, -DPB1, MLC negative (Table 4). Ten out of these 12 patients survived and are remarkably well (manuscript in preparation). Length of time between search request and transplantation Less than 1 day was needed to identify a potential donor in Europdonor due to the computer register being on site. For the other registries, where transmission of requests and results was necessary, this interval decreased between 1990 and 1995 from a median of 11 days to 4 days. The time needed to obtain a blood sample from a potential donor differed substantially between registries but was a median of 9 days (range 1–38) for donors from the Europdonor file and 19 days (range 3–149) from elsewhere. It should be

Finding an unrelated bone marrow donor M Oudshoorn et al

Table 4

Degree of matching in 75 transplanted patient/donor pairsa

HLA match category

A B DR (serology, split) A B DRB1/3/4/5 DQB1 A B C DRB1/3/4/5 DQB1 A B C DRB1/3/4/5 DQB1 DPB1 A B C DRB1/3/4/5 DQB1 DPB1 MLC negative A B C DRB1/3/4/5 DQB1 DPB1 MLC negative, CTLp negative

No. of pairs matched/ Total No. of pairs tested 65/75 60/75 42/75 12/75 12/75

= = = = =

87% 80% 56% 16% 16%

4/54b = 7%

a

For the remaining 11 transplanted patient/donor pairs no material was left to perform retrospective typing at DNA level. b The CTLp test was only performed successfully in 54 patient/donor pairs.

noted that samples from other registries were obtained much faster in 1993–1995 than before 1993. The time needed for donor identification and the time to transplantation was substantially shorter for donors residing in The Netherlands compared to international donors; 3.0 vs 5.1 months for donor choice and 5.1 vs 7.0 months for time to transplantation. It is unclear to what extent this difference in time interval is caused by the fact that for patients with a common phenotype there will be several potential donors in the national file. Search outcome for patients from non-European origin The percentage of patients of non-European origin rose from 5% in 1990 to 16% in 1995. In total, 26 of the 240 patients were of non-European origin. For seven patients (27%) an HLA-A,-B,-DR split identical donor could be identified in BMDW. Only four out of the 26 patients, of whom two were an identical twin, were transplanted. For the remaining three patients with a potential donor, the search was stopped; one patient for personal reasons, for one patient HLA subtyping of the donor revealed too many mismatches and for one patient due to advanced disease status. Discussion This study has been useful in pinpointing some of the obstacles encountered when searching for an HLA matched unrelated donor. The analysis of the 240 searches showed that in 1995 for over 70% of the patients an HLA-A,-B,-DR split identical donor was available among the more than 500 000 split typed donors reported in BMDW. This is approximately 10% higher than predicted by Beatty et al23 in 1988. Less than 20% of the total number of donors reported in BMDW is HLA-A,-B,-DR split typed but 86 out of 87 (99%) patients transplanted received a graft from an already HLAA,-B,-DR typed donor. Similar findings have been observed in the file of the NMDP (P Coppo, personal communication). One might conclude that the best way to proceed is not to enlarge the registries further by recruiting new donors but in contrast, by retyping available donors

more precisely and including HLA-DR typing as also suggested by Speiser et al.11 There is a large difference between the percentage of patients for whom a donor seems to be available (50% in 1990–1991 and 79% in 1994–1995) and the percentage of patients actually transplanted (24% in 1990–1991 and 44% in 1994–1995), even though in this study the percentage of patients transplanted is higher than reported elsewhere.8,10,24 For 8% of patients this is due to donor withdrawal (4%) or incorrect typing of the donor (4%). A disturbing trend in this respect is the increase of donors (7% in 1990 vs 17% in 1995) who are unavailable mainly due to withdrawal of the donor for personal reasons. According to a recent analysis by the National Marrow Donor Program blood transfusion donors and apheresis donors have a lower unavailability rate (16.7%) than donors recruited by community drives (26.4%). In Europdonor, where mostly blood transfusion donors who have donated at least twice are recruited as bone marrow donors, no increase in donor withdrawal rate is seen over the years (9% in 1990 and 10% in 1995). The overall percentage of donors with HLA typing errors is 8% but varies between the different registries from less than 1% to over 25%. Homozygosity either missed or incorrectly assumed to be present, was the cause of almost half of the mistypings. It is thus advisable to request a higher number of donors for testing when the patient is homozygous for one or more loci. The inclusion of the MLC and CTLp test in our protocol deserves a comment. Several publications stress that an MLC test is often uninterpretable and taken alone is not a predictor of a better transplant prognosis.25–27 We agree that the outcome of the MLC test alone is not a useful criterion for donor selection. We still perform the MLC test as a control for HLA-DRB1 typing as new alleles may not be detected by the PCR-based typing technique used, and to answer the question whether it is a predictor of GVHD or relapse in combination with HLA-DPB1. The outcome of the CTLp test is useful for selecting the best donor as has been documented by others.28–30 It recognizes mismatches for class I, including HLA-C alleles, and serves as a check for the correctness of HLA class I typing. Theoretically, it could recognize minor histocompatibility mismatches but we have not encountered that.31 Thirteen percent of patients (n = 31) were withdrawn during the search process for reasons other than the unavailability of an unrelated donor. The reasons (n = 20) were therapy-resistant leukaemia, unwillingness of the patient to be transplanted (n = 6), or the identification of a suitable related donor after an extended family search (n = 5). For 17% of patients for whom an unrelated donor was available the patient may have been referred to the transplant centre too late or the search process took too long. They were either deceased or were no longer transplantable due to advanced disease status. This ought to be an incentive to reorganize the search process drastically in order to shorten the time needed to identify a suitable donor. It has been shown that this can, if necessary, be reduced to 1 week in at least some cases (Ref. 32, and Cleaver, personal communications). An additional reason why some patients with an HLA-A,-B,-DR serologically typed donor are not

1015

Finding an unrelated bone marrow donor M Oudshoorn et al

1016

transplanted is the presence of too many mismatches detected at DNA level. For 89% of patients with at least one HLA-A,-B,-DR and -DQ identical donor as determined by serology, one (or more) HLA-A,-B,-DRB1,-DQB1 donor(s) were identified. This figure is much higher than is reported in the literature.11 One reason may be that we, whenever possible, select donors beforehand who are also identical for HLA-C and -DQ. Testing for DRB3,4 or 5 did not change the numbers while DPB1 was only matched in 28% of the patients who were HLA-A,-B,-DRB1/3/4/5,-DQB1 identical. This figure is higher than reported elsewhere,9,12,33 possibly because of a larger number of donors tested per patient by us (mean 6.7). These figures were reduced markedly for donors with either an HLA-A or -B locus mismatch. In this situation only 50% of patients had at least one HLA-A or -B, -DRB1/3/4/5, -DQB1 identical donor and 9% of these were also DPB1 identical. This is probably a reflection of linkage disequilibrium. We can only speculate why patients with one haplotype in linkage disequilibrium (F) and one haplotype without (N) appear to do almost equally well in finding a donor as patients with two haplotypes in linkage disequilibrium. Most probably this is due to the fact that the chance of finding a donor is directly correlated to the product of the frequencies of the two most common haplotypes out of the six possible. If one haplotype is very frequent and the number of donors is large, we can equal the chance to one. In other words, finding a donor with that haplotype is not restricted. As a result, the frequency of the less frequent haplotype is the only limiting factor. The number of haplotypes is still very large, but only a square root function of the phenotypes, and apparently does not function as a restraint if a large donor pool is available. For 81 patients (34%) no donor was available. Of this group, 19 patients originated from Morocco, Turkey or (former) Dutch colonies. In general, the search outcome for 26 patients of non-European origin was unsatisfactory as expected; only four patients, of whom two were identical twins, were transplanted. This is in accordance with the data of Beatty et al34 and the IMUST study.8 There is an increase of patients of non-European origin (5% in 1990 to 16% in 1995) reported to Europdonor and this has an effect on the search outcome. The slight decrease of available donors in 1995 (72%) compared to 1994 (75%), despite the increase of potential donors, is due to the higher percentage of patients of non-European origin. These patients, children in particular, may be better helped by increasing the availability of cord blood units from non-European populations.35,36 While writing this report we were confronted with the lack of a generally accepted term to indicate the extent to which donor and recipient are matched. This makes it difficult to compare the success rate in locating a ‘matched’ donor as reported in the different publications. It is often not clear whether donor and recipient are ‘identical’ on the basis of the serologically defined broad HLA antigens or whether splits were also tested for. The same holds true for the use of the different DNA-based typing techniques. We would like to propose that in analogy to the grading of the severity of acute graft-versus-host disease in five classes37

an agreement is reached for the different matching categories. It is obvious that this should be discussed and agreed by the World Marrow Donor Association (WMDA). Finally, we should like to re-emphasize that in the complicated process of finding and selecting an unrelated donor and obtaining, transporting and transplanting the bone marrow, intensive interaction between the registry staff, the immunogeneticist, immunologist and clinician is essential.32 It is only in this way that the patient can be optimally helped and that opportunities such as finding a suitable family donor outside the core family through an extended family search are not missed. Acknowledgements This work has been supported in part by the Immunology Working Party from the European Group for Blood and Marrow Transplantation (EBMT) (EC Grant BMHI-CT 94–0300), the JA Cohen Institute for Radiopathology and Radiation Protection (IRS), the Dutch National Reference Centre for Histocompatibility, the World Marrow Donor Association, the Nijbakker Morra Foundation, the Roche Diagnostic Systems and Mrs J Baaij. We would like to thank Dr A Brand and Professor FHJ Claas for critically reviewing the manuscript and Professor H Kolb for useful suggestions.

References 1 Stroncek D. Results of bone marrow transplants from unrelated donors. Transfusion 1992; 32: 180–189. 2 Kernan NA, Bartsch G, Ash RC et al. Analysis of 462 transplantations from unrelated donors facilitated by the National Marrow Donor Program. New Engl J Med 1993; 328: 595– 602. 3 Van Rood JJ, van Leeuwen A, van den Tweel JG et al. Typing for the MLC determinants and its possible impact on Europdonor. Transplant Proc 1974; VI: 411–415. 4 McCullough J, Rogers G, Dahl R et al. Development and operation of a program to obtain volunteer bone marrow donors unrelated to the patient. Transfusion 1986; 26: 315– 323. 5 McCullough J, Hansen JA, Perkins HA et al. The NMDP: how it works, accomplishments to date. Oncology 1989; 3: 63–72. 6 Cleaver SA. The Anthony Nolan Research Centre. Bone Marrow Transplant 1993; 11 (Suppl. 1): 38–40. 7 Stroncek D, Bartsch G, Perkins HA et al. The National Marrow Donor Program. Transfusion 1993; 33: 567–577. 8 Howard MR, Gore SM, Hows JM et al. A prospective study of factors determining the outcome of unrelated marrow donor searches: report from the International Marrow Unrelated Search and Transplant Study Working Group on behalf of collaborating centres. Bone Marrow Transplant 1994; 13: 389– 396. 9 Anasetti C, Howe C, Petersdorf EW et al. Marrow transplants from HLA matched unrelated donors: an NMDP update and the Seattle experience. Bone Marrow Transplant 1994; 13: 693–695. 10 Ottinger H, Grosse-Wilde M, Schmitz A, Grosse-Wilde H. Immunogenetic marrow donor search for 1012 patients: a retrospective analysis of strategies, outcome and costs. Bone Marrow Transplant 1994; 14 (Suppl. 4): S34–S38. 11 Speiser DE, Tiercy J-M, Rufer N et al. Relation between the

Finding an unrelated bone marrow donor M Oudshoorn et al

12

13 14

15 16

17

18

19

20 21 22 23 24

resolution of HLA-typing and the chance of finding an unrelated bone marrow donor. Bone Marrow Transplant 1994; 13: 805–809. Hansen JA, Petersdorf EW, Yoon Choo S et al. Marrow transplantation from HLA partially matched relatives and unrelated donors. Bone Marrow Transplant 1995; 15 (Suppl. 1): S128–S139. Schipper RF, D’Amaro J, Oudshoorn M. The probability of finding a suitable related donor for bone marrow transplant in extended families. Blood 1996; 87: 800–804. Oudshoorn M, van Leeuwen A, van der Zanden HGM, van Rood JJ on behalf of the Editorial Board of BMDW. Bone Marrow Donors Worldwide. A successful exercise in international cooperation. Bone Marrow Transplant 1994; 14: 3–8. Schipper RF, D’Amaro J, Oudshoorn M. The probability of finding a haplotypically identical unrelated bone marrow donor. Hum Immunol 1996 (submitted). Van Rood JJ, van Leeuwen A, Ploem JS. Simultaneous detection of two cell populations by two-colour fluorescence and application to the recognition of B-cell determinants. Nature 1976; 262: 795–797. Verduyn W, Doxiadis IIN, Anholts JDH et al. Biotinylated DRB sequence specific oligonucleotides, comparison to serologic HLA-DR typing of organ donors in Eurotransplant. Hum Immunol 1993; 37: 59–67. Olerup O, Zetterquist H. HLA-DR typing by 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 Antigen 1993; 39: 225–235. Bunce M, O’Neill CM, Barnardo MCNM et al. Phenotyping: comprehensive DNA typing for HLA-A, B, C, DRB1, DRB3, DRB4, DRB5 & DQB1 by PCR with 144 primer mixes utilising sequence-specific primers (PCR-SSP). Tissue Antigen 1995; 46: 355–367. Keuning JJ, Termijtelen A, Blusse´ van Oud Alblas A et al. Typing for MLC (LD). Transplant Proc 1975; VII: 35–39. Zhang L, Li S, Vandekerckhove B et al. Analysis of cytotoxic T cell precursor frequencies directed against individual HLAA and -B alloantigens. J Immunol Meth 1989; 121: 39–45. Schipper RF, Schreuder GMTh, D’Amaro J, Oudshoorn M. HLA and gene and haplotype frequencies in Dutch blood donors. Tissue Antigen 1996; 48: 562–574. Beatty PG, Dahlberg S, Mickelson EM et al. Probability of finding HLA-matched unrelated marrow donors. Transplantation 1988; 45: 714–718. Dini G, Lanino E, Lamparelli T et al. Unrelated donor marrow transplantation: initial experience of the Italian Bone Marrow

25

26

27

28

29

30

31 32

33

34 35 36 37

Transplant Group (GITMO). Bone Marrow Transplant 1996; 17: 55–62. Mickelson EM, Bartsch GE, Hansen JA et al. The MLC assay as a test for HLA-D region compatibility between patients and unrelated donors: results of a national marrow donor program involving multiple centers. Tissue Antigen 1993; 42: 465–472. Segall M, Noreen H, Edwins L et al. Lack of correlation of MLC reactivity with acute graft-versus-host disease and mortality in unrelated donor bone marrow transplantation. Hum Immunol 1996; 49: 49–55. Mickelson EM, Longton G, Anasetti C et al. Evaluation of the mixed lymphocyte culture (MLC) assay as a method for selecting unrelated donors for marrow transplantation. Tissue Antigen 1996; 47: 27–36. Kaminski E, Hows J, Man S et al. Prediction of graft versus host disease by frequency analysis of cytotoxic T cells after unrelated donor bone marrow transplantation. Transplantation 1989; 48: 608–613. Spencer A, Brookes PA, Kaminski E et al. Cytotoxic T lymphocyte precursor frequency analysis in bone marrow transplantation with volunteer unrelated donors. Transplantation 1995; 59: 1302–1308. Speiser DE, Lo¨liger C-C, Siren M-K et al. Pretransplant cytotoxic donor T-cell activity specific to patient HLA class I antigens correlating with mortality after unrelated BMT. Br J Haematol 1996; 93: 935–939. Rufer N, Tiercy J-M, Breur-Vriesendorp B et al. Histoincompatibilities in ABDR-matched unrelated donor recipient combinations. Bone Marrow Transplant 1995; 16: 641–646. Oudshoorn M, Falkenburg JHF, Ebeling LJ et al. Case Report: Unrelated bone marrow transplantation as a rescue procedure following inadvertent loss of an autologous bone marrow graft. Bone Marrow Transplant 1996; 18: 461–463. Petersdorf EW, Perkins H, Sekiguchi S et al. Joint Report. Marrow transplantation. In: Tsuji K, Aizawa M, Sasazuki T (eds). HLA 1991, Vol. I. Oxford University Press: Oxford, 1992, pp 812–820. Beatty PG, Mori M, Milford E. Impact of racial genetic polymorphism on the probability of finding an HLA-matched donor. Transplantation 1995; 60: 778–783. Wagner JE, Kernan NA, Steinbuch M et al. Allogeneic sibling umbilical-cord-blood transplantation in children with malignant and non-malignant disease. Lancet 1995; 346: 214–219. Kurtzberg J, Laughlin M, Graham ML et al. Placental blood as a source of hematopoietic stem cells for transplantation into unrelated recipients. New Engl J Med 1996; 335: 157–166. Gale RP, Mortin MM, van Bekkum DW et al. Risk factors for acute graft-vs-host disease. Br J Haematol 1987; 67: 397–406.

1017