Methods

Top Methods Statistical Methods Results Discussion Author & Article Info References

The study population included 103 patients with ALL or AML in first remission or CML in first chronic phase receiving bone marrow transplants from identical twins between 1978 and 1990 and reported to the International Bone Marrow Transplant Registry by 66 centers. During the same period, 3214 recipients of non-T-cell-depleted HLA-identical sibling transplants for these indications were reported to the International Bone Marrow Transplant Registry by 163 centers: 581 for ALL in first remission, 1394 for AML in first remission, and 1240 for CML in first chronic phase.

Assessing Comparability of Twin and HLA-identical Sibling Transplant Recipients

For each disease, prognostic factors for transplant outcome were compared between twin and HLA-identical sibling transplant recipients using chi-square for categorical variables and the Mann-Whitney test [2] for continuous variables. Factors considered were those associated with HLA-identical sibling transplant outcome in previous International Bone Marrow Transplant Registry studies [3-7]. These included the following: for ALL, age, leukocytes at diagnosis, immune phenotype, and time to achieve first remission [3, 4]; for AML, age, leukocytes at diagnosis, FAB type, and Karnofsky performance score [5]; and for CML, age, previous splenectomy, interval between diagnosis and transplant, and previous treatment with busulfan [6, 7]. Conditioning regimens were also compared.

For each twin transplant, 10 controls were selected from among 3214 HLA-identical sibling transplants matched for disease and any prognostic factors differing between the two cohorts with P < 0.1. For ALL, pairs were matched for age within 5 years; for AML, pairs were matched for FAB type and Karnofsky score; for CML, pairs were matched for age, interval between diagnosis and transplant, and use of busulfan before transplantation.

Statistical Methods

Top Methods Statistical Methods Results Discussion Author & Article Info References

Actuarial probabilities of relapse, treatment-related mortality [8], and leukemia-free survival were calculated in the twins and matched HLA-identical sibling controls using life-table methods. Relapse was defined as hematologic recurrence in any site; patients in continuous complete remission were censored at death or, for survivors, at last contact. Treatment-related mortality was defined as death in continuous complete remission; patients were censored at time of relapse or at last follow-up. Leukemia-free survival was defined as survival in continuous complete remission. Relapse and death from causes other than leukemia were considered failures; patients alive and in remission were censored at time of last follow-up. Univariate comparisons of survival distributions were done using the matched log-rank test [9]. To compare relapse risks after adjusting for the effect of graft-versus-host disease, we used Cox proportional-hazards regression with acute and chronic graft-versus-host disease entered as time-dependent covariates [10, 11].

The International Bone Marrow Transplant Registry

The Registry is a voluntary working group of more than 200 transplant teams worldwide that contribute detailed data on their allogeneic and identical-twin bone marrow transplants to a statistical center [12, 13]. The program is primarily funded by a program project grant from the U.S. National Institutes of Health. Participants are required to report all consecutive transplants; compliance is monitored by on-site audits. Approximately two thirds of all active allogeneic transplant centers report their data to the Registry. The Registry database includes 40% to 45% of all allogeneic transplant recipients since 1970. Patients are followed longitudinally. Computerized error checks, physician review of submitted data, and on-site audits of participating centers ensure data quality. Transplant outcomes estimated using Registry data are similar to those reported by large nonparticipating centers for comparable patients.

Results

Top Methods Statistical Methods Results Discussion Author & Article Info References

Differences were found in patient-, disease- and treatment-related variables between the 103 recipients of identical-twin transplants and the 3214 recipients of HLA-identical sibling transplants reported to the International Bone Marrow Transplant Registry during the same interval. Some were intrinsic to the study design. For example, identical-twin donors and recipients were always of the same sex compared with only about one half of HLA-identical sibling donor-recipient pairs. Also, only eight (8%) twin recipients received immune suppression after transplantation compared with 100% of HLA-identical sibling transplant recipients. It was not possible to match or otherwise control for potential effects of these differences. Other differences were as follows: for ALL, younger median age in twins compared with HLA-identical siblings (17 compared with 22 years; P = 0.003); for AML, higher proportion of twins older than 40 years of age (22% compared with 8%; P = 0.002) and with Karnofsky performance scores at transplantation of less than 90% (15% compared with 10%; P < 0.001); and for CML in first chronic phase, shorter median interval from diagnosis to transplant in twins (6 compared with 13 months; P < 0.001) and fewer twins treated with busulfan before transplant (12% compared with 32%, P = 0.04). Ten HLA-identical sibling transplant controls were selected for each twin transplant, matching for diagnosis and for these variables. The resulting matched cohort was similar to the twin cohort for all variables considered, other than those intrinsic to twin transplants as indicated (Appendix Table 1). The median follow-up times of 6.3 years (range, 0.4 to 13.6 years) and 5.0 years (range, 0.3 to 14.8 years) were similar for twin and HLA-identical sibling transplant cohorts, respectively (P = 0.2).

Transplant Outcome

Acute Lymphoblastic Leukemia

Ten of 24 twins had a relapse. The 3-year probability of relapse was 36% (CI, 17% to 55%) compared with 26% (CI, 20% to 32%) after 240 HLA-identical sibling transplants (P = 0.1; Figure 1). The 95% CI for the difference in relapse rates between twin and HLA-identical sibling transplants at 3 years was –10% to 30%. Regression analysis with adjustment for acute and chronic graft-versus-host disease showed no difference between identical-twin and HLA-identical sibling transplants in relapse risk (relative risk for twins compared with HLA-identical siblings, 1.4; CI, 0.6 to 2.8, P > 0.2).

View larger version (16K): [in this window] [in a new window]   Figure 1. Outcome of transplants for acute lymphoblastic leukemia. Actuarial probability of relapse (top) and leukemia-free survival (bottom) after identical-twin and HLA-identical sibling bone marrow transplants for acute lymphoblastic leukemia in first remission. Numbers in parentheses are numbers at risk (alive in continuous complete remission) at indicated intervals.

Acute Myelogenous Leukemia

Twenty-three of 45 twins had a relapse. The 3-year probability of relapse was 52% (CI, 37% to 67%) compared with 16% (CI, 12% to 20%) after 450 HLA-identical sibling transplants (P < 0.001; Figure 2). The 95% CI for the difference in relapse rates at 3 years was 20% to 52%. Regression analysis after adjustment for acute and chronic graft-versus-host disease also showed an increased relapse risk in twins (relative risk, 3.8; CI, 2.3 to 6.2, P < 0.001).

View larger version (26K): [in this window] [in a new window]   Figure 2. Outcome of transplants for acute myelogenous leukemia. Actuarial probability of relapse (top) and leukemia-free survival (bottom) after identical-twin and HLA-identical sibling bone marrow transplants for acute myelogenous leukemia in first remission. Numbers in parentheses are numbers at risk (alive in continuous complete remission) at indicated intervals.

Chronic Myelogenous Leukemia

Seventeen of 34 twins had a relapse. The 3-year probability of relapse was 40% (CI, 23% to 57%) compared with 7% (CI, 4% to 10%) after 340 HLA-identical sibling transplants (P < 0.001; Figure 3). The 95% CI for the difference in relapse rates at 3 years was 16% to 50%. Regression analysis after adjustment for acute and chronic graft-versus-host disease also showed increased relapse risk in twins (relative risk, 5.5; CI, 2.8 to 11.0, P < 0.001).

View larger version (26K): [in this window] [in a new window]   Figure 3. Outcome of transplants for chronic myelogenous leukemia. Actuarial probability of relapse (top) and leukemia-free survival (bottom) after identical-twin and HLA-identical sibling bone marrow transplantation for chronic myelogenous leukemia in chronic phase. Numbers in parentheses are numbers at risk (alive in continuous complete remission) at indicated intervals.

Discussion

Top Methods Statistical Methods Results Discussion Author & Article Info References

Our analysis indicates that in AML and CML, persons receiving transplants from identical twins have an increase in relapse that is statistically significant when compared with similar persons receiving transplants from HLA-identical siblings. Although a similar trend was observed in ALL, it was not statistically significant. This difference persists after adjusting for effects of acute and chronic graft-versus-host disease. Leukemia-free survival is similar after twin and HLA-identical sibling transplants in all three types of leukemia.

The finding that persons receiving twin transplants for AML and CML have an increased risk for leukemia relapse compared with HLA-identical sibling transplant recipients is not surprising in view of reports of an antileukemia effect of graft-versus-host disease (for reviews, see 14-19), which does not generally occur after twin transplants. However, in our study, the increased relapse risk for twins persisted after adjusting for graft-versus-host disease. Which, if any, additional antileukemia mechanisms, such as subclinical graft-versus-host disease or graft-versus-leukemia, operate in the context of genetic disparity between donor and recipient is unknown.

In contrast to AML and CML, the increased relapse risk in recipients of twin transplants for ALL was less impressive and not statistically significant when compared with HLA-identical sibling transplant recipients. This is consistent with our previous analyses [17, 18]. The reason for this difference between the myeloid leukemias and ALL is unknown. One possibility is that the smaller number of patients with ALL than with AML or CML limited the power of the analysis. The 95% CI for the difference in relapse rates was wide, and it is possible that this study missed a real difference of as much as 30% (type II error). It is also possible that a substantial proportion of persons with ALL receiving transplants in first remission are cured by chemotherapy before receiving their transplant. Consequently, decreased antileukemia effects might have little or no effect. A third possibility is that intensive pretransplant conditioning is more effective in ALL than in AML or CML. This is consistent with chemotherapy studies indicating a more convincing dose-response relation for ALL than myeloid leukemias (reviewed in 20). Finally, graft-versus-leukemia effects may not operate in ALL.

Several aspects of our study deserve further discussion. One is that only 103 twin transplants were available for analysis. Although this number limits the power of the analysis, we estimate that these data represent about one half of all comparable twin cases worldwide. Further, in many instances we were able to detect differences in transplant outcomes despite the small number of patients analyzed.

One possible explanation for an increased relapse risk in twins with AML and CML might have been differences in pretransplant risk factors. However, the twin and HLA-identical sibling cohorts were fairly comparable, and none of the few differences we identified (AML, performance score; CML, interval from diagnosis to transplant and previous busulfan treatment) have had a substantial effect on relapse risk in our previous analyses of HLA-identical siblings. Further, an increased relapse risk was detected after matching for these factors. It is unlikely, therefore, that the increased relapse risk results from other differences between twin and HLA-identical sibling transplant recipients.

One issue we were unable to address was the possibility of differences in cytogenetic abnormalities in patients with AML. However, this is also unlikely to explain differences in relapse rates. In a previous study of more than 700 recipients of HLA-identical sibling transplants for AML, we found an increased relapse risk only in persons with –5,-7, or t(9; 22) abnormalities [21]. All others, including those without cytogenetic abnormalities, had comparable relapse risks. This adverse prognosis group accounted for less than 5% of all cases. In our study, only 1 of 18 evaluable twins with AML had an adverse cytogenetic abnormality, t(9; 22). By definition, no patients with CML in chronic phase had additional cytogenetic abnormalities because these patients would have been classified as being in an accelerated phase. Thus, differences in cytogenetic features do not account for the increased relapse risk in twins with AML and CML.

A third issue is related to reports of acute graft-versus-host disease after twin transplants [22]. It is uncertain whether these cases are correctly diagnosed. We previously emphasized the considerable difficulty in accurately diagnosing acute graft-versus-host disease [23] and questioned whether this diagnosis is correct in twins [24]. Another factor is that graft-versus-host disease reported in twins is usually mild and would not be expected to be associated with a substantial antileukemia effect based on parallel data in allograft recipients [20]. Finally, whether or not twins develop acute graft-versus-host disease, it need not correlate with the same antileukemia effect observed after HLA-identical sibling transplants.

The risk for relapse in persons in chronic phase CML in our study is similar to that reported in a study of twins receiving transplants with comparable pretransplant conditioning but higher than that reported in eight twins receiving dimethylmyleran before transplantation [25]. We were unable to address this issue because no one in our data set received dimethylmyleran.

Although twin transplants were associated with an increased risk for relapse in AML and CML, leukemia-free survival was similar to that for recipients of HLA-identical sibling transplants because increased relapses were offset by less treatment-related mortality.

Analysis of relapse rates after twin compared with HLA-identical sibling transplants is useful to predict whether high doses of chemotherapy and radiation are likely to cure leukemia in the absence of graft-versus-host disease and other possible immune-mediated antileukemia effects, a question important in predicting the outcome of more intensive chemotherapy and autotransplants. Our prediction that relapse risks after autotransplant are likely to be high in AML in first remission and CML in chronic phase is supported by published data [26, 27]. The high relapse rates reported after autotransplants for ALL contrast with the relapse rates after twin transplants and suggest that leukemia cells in the graft may contribute to relapse.

Twin transplant data also indicate what results can be expected if immunologic problems, like graft-versus-host disease, and confounding issues like immune deficiency and interstitial pneumonia, were eliminated or controlled after HLA-identical sibling transplants. Our data suggest that the intrinsic failure rate (treatment-related mortality) for patients with transplants for acute leukemia in first remission or CML in first chronic phase is between 5% and 10%.

Transplants between identical twins are rare, accounting for fewer than 2% of the transplants reported between 1980 and 1990 [1]. Because the results of these transplants are important in understanding transplant biology, we recommend reporting all cases to the International Bone Marrow Transplant Registry.

Participating institutions: Transplant centers contributing data to this analysis included the following: Argentina: British Hospital of Buenos Aires, Buenos Aires; Navy Hospital "Pedro Mallo," Buenos Aires. Australia: Hanson Center for Cancer Research, Adelaide; Royal Brisbane Hospital, Brisbane; Royal Prince Alfred Hospital, Camperdown; St. Vincent's Hospital, Darlinghurst; Royal Hobart Hospital, Hobart; Royal Melbourne Hospital, Parkville; Royal Perth Hospital, Perth; Alfred Hospital, Prahran; Prince of Wales Children's Hospital, Randwick; Westmead Centre, Westmead; Queen Elizabeth Hospital, Woodville. Austria: Med. Univ.-Klinik, Vienna. Belgium: Cliniques Univ. Saint-Luc, Bruxelles; University of Leuven, Leuven. Brazil: Instituto Nacional de Cancer, Rio de Janeiro; Hospital de Clinicas, Curitiba. Canada: Tom Baker Cancer Centre, Calgary; Chedokee-McMaster University, Hamilton; Royal Victoria Hospital, Montreal; Montreal Children's Hospital, Montreal; Princess Margaret Hospital, Toronto; Vancouver General Hospital, Vancouver. China: Beijing Medical University, Beijing; Bei Tai Ping Lu Hospital, Beijing; Lanzhou General Hospital, Lanzhou. Croatia: Klinika za Unutrasnje bolesti KBC-Rebro, Zagreb. Denmark: Rigshospitalet, Copenhagen. England: East Birmingham Hospital, Birmingham; Queen Elizabeth Medical Center, Birmingham; St. James's University Hospital, Leeds; Charing Cross & Westminster Medical School, London; The London Clinic, London; Royal Free Hospital, London; Royal Marsden Hospital, London; Royal Postgraduate Medical School, London; Westminster Children's Hospital, London; Royal Victoria Infirmary, Newcastle; John Radcliffe Hospital, Oxford; London Hospital Whitechapel, Whitechapel. Finland: Helsinki University Central Hospital, Helsinki; Turku University, Turku. France: Angers University, Angers; Hopital Jean-Minjoz, Besancon; Universitaire de Caen, Caen; Hopital A. Michallon, Grenoble; Centre Hospital Regional De Lille, Lille; Hopital Debrousse, Lyon; Institut J. Paoli I. Calmettes, Marseille; Centre Hospitalier Regional, Nantes; Hotel Dieu de Paris, Paris; Hopital Saint-Antoine, Paris; Hopital Saint-Louis, Paris; Groupe Hospitalier du Hau Leveque, Pessac; Hopital Jean Bernard, Poitiers; Hopital Nord, St. Etienne; Hopital Purpan, Toulouse; Hospital Regional de Toulouse, Toulouse. Germany: University of Leipzig, Leipzig; Universitatsklinikum Charlottenburg, Berlin; Medizinische Hochschule, Hannover; Christian-Albrechts-Universitat, Kiel; Universitats Kinderklinik, Munich; Universitat Muenchen, Munich; Med Universitatsklinik, Tubingen; Universitat Ulm, Ulm/Donau. Hungary: National Institute of Haematology, Budapest; Semmelweis University, Budapest. India: Tata Memorial Hospital, Bombay. Ireland: St. James's Hospital, Dublin. Italy: Spedali Civili, Brescia; S. Orsola University Hospital, Bologna; Ospedale San Martino, Genoa; University of Milan, Milan; Ospedale Civile, Pesaro; Centro Trapianti Midollo Osseo, Pescara; Ospedale S. Camillo, Rome; University degli Studi, La Sapienza, Rome. Japan: Tokai University School of Medicine, Kanazawa; Kanazawa University School of Medicine, Kanazawa-shi; Nagoya Second Red Cross Hospital, Nagoya; Center For Adult Diseases, Osaka; Hyogo College of Medicine, Hyogo; University of Tokyo, Tokyo. Korea: St. Mary's Hospital, Seoul. Malaysia: University of Malay, Kuala Lumpur. Netherlands: Leiden University Hospital, Leiden; University of Nijmegen, Nijmegen; Dr. Daniel den Hoed Cancer Center, Rotterdam. New Zealand: Auckland Hospital, Auckland; Christchurch Hospital, Christchurch; Wellington School of Medicine, Wellington. Norway: Rikshospitalet, Oslo. Poland: Postgraduate Medical Center, Warsaw. Portugal: Instituto Portugues, Lisbon. Russia: Ministry of Health for USSR, Moscow. Saudi Arabia: King Faisal Specialist Hospital, Riyadh; Riyadh Armed Forces Hospital, Riyadh. Scotland: Royal Infirmary of Edinburgh, Edinburgh; Glasgow Royal Infirmary, Glasgow. South Africa: University of Cape Town Medical School, Cape Town; University of Witwatersrand, Johannesburg. Spain: Hospital General Vall d'Hebron, Barcelona; Hospital Infantil Vall d'Hebron, Barcelona; University of Barcelona, Barcelona; Hospital "Reina Sofia," Cordoba; Clinica Puerta de Hierro, Madrid; Hospital de la Princesa, Madrid; Hospital Ramon y Cajal, Madrid; Hospital "M. De Valdecilla," Santander; Hospital "La Fe," Valencia. Sweden: Huddinge Hospital, Huddinge; University of Goteborg, Goteborg; University of Lund, Lund; Switzerland: Kantonsspital, Basel; Kindenspital, Zurich. Taiwan: National Taiwan University Hospital, Taipei; Veterans General Hospital, Taipei. United States: Emory University, Atlanta, Georgia; Johns Hopkins Oncology Center, Baltimore, Maryland; Alta Bates Hospital, Berkeley, California; University of Alabama, Birmingham, Alabama; Roswell Park Memorial Institute, Buffalo, New York; University of Virginia, Charlottesville, Virginia; Rush-Presbyterian, Chicago, Illinois; Children's Hospital, Cincinnati, Ohio; Case Western Reserve University, Cleveland, Ohio; Cleveland Clinic, Cleveland, Ohio; City of Hope, Duarte, California; University of Florida, Gainesville, Florida; Children's Hospital, Houston, Texas; St. Joseph's Hospital Medical Center, Houston, Texas; M. D. Anderson Hospital, Houston, Texas; Indiana University Hospital, Indianapolis, Indiana; University of Kansas, Kansas City, Kansas; Wilford Hall Medical Center, Lackland AFB, California; University of Kentucky, Lexington, Kentucky; Southern California Permanente Medical Group, Los Angeles, California; University of California/Medicine, Los Angeles, California; James Graham Brown Cancer Center, Louisville; University of Wisconsin/Medicine, Madison, Wisconsin; Marshfield Clinic, Marshfield, Wisconsin; Loyola University, Maywood, Illinois; Medical College of Wisconsin, Milwaukee, Wisconsin; University of Minnesota, Minneapolis, Minnesota; Louisiana State Medical Center, New Orleans, Louisiana; Memorial Sloan-Kettering, New York, New York; University of Oklahoma, Oklahoma City, Oklahoma; Bishop Clarkson Memorial Hospital, Omaha, Nebraska; University of Nebraska, Omaha, Nebraska; The Children's Hospital, Philadelphia, Pennsylvania; Hahnemann University, Philadelphia, Pennsylvania; Montefiore Hospital, Pittsburgh, Pennsylvania; LDS Hospital, Salt Lake City, Utah; Moffit Hospital, San Francisco, California; Stanford University Hospital, Stanford, California; Georgetown University Medical Center, Washington, DC. Venezuela: Hospital Central de Valencia, Valencia. Wales: University Hospital of Wales, Cardiff.