Transplant Registries: Guiding Clinical Decisions and Improving Outcomes

ABSTRACT: About 50,000 hematopoietic stem cell transplantations areperformed yearly, primarily for malignancies. Use of this therapy increaseddramatically over the past 30 years due to its proven and potential efficacy indiverse diseases, better understanding of appropriate timing of transplantationand patient selection, and greater availability of allogeneic donors. TheInternational Bone Marrow Transplant Registry (IBMTR) and the Autologous Bloodand Marrow Transplant Registry (ABMTR) collect data on consecutive allogeneicand autologous transplants, respectively, in more than 400 participating centersworldwide. The IBMTR/ABMTR database contains information on more than 120,000transplant recipients. Among 11,347 patients transplanted in 101 IBMTR/ABMTRresearch centers in North America during 1995-1997, 66% received autologoustransplants, 24% related-donor transplants, and 10% unrelated-donor transplants.More than 90% of transplantations were for malignant disease, with more thanhalf of these done in patients with advanced disease. Of the recipients, 70%were younger than 50 years. Posttransplant survivals varied substantially bydisease, transplant type, recipient age, and disease status at transplantation.IBMTR/ABMTR data provide an important tool for assessing transplant use andoutcome, identifying prognostic factors for transplant outcomes, evaluating newtransplant therapies, comparing transplant and nontransplant therapies,evaluating late transplant complications, and planning prospective phase II andIII clinical trials. [ONCOLOGY 15(5):649-666, 2001]

Introduction

The first successful transplants ofhematopoietic stem cells wereperformed in 1968 in three children with congenital immune deficiencydiseases.[1-4] In each instance, stem cells were collected from the bone marrowof sibling donors who were genotypically identical or closely matched to therecipient for human leukocyte antigens (HLA). Since then, thousands of patientshave received hematopoietic stem cell transplants as treatment for malignant andnonmalignant diseases.

Approximately 50,000 transplants are performed worldwide eachyear (Figure 1). Reasons for the increased use of stem cell transplantation overthe past 3 decades include its proven and potential efficacy in many diseases,better understanding of the appropriate timing of transplantation and patientselection, greater availability of donors, better techniques for determining HLAmatch, greater ease of stem cell collection, and improved supportive careresulting in less transplant-related morbidity and mortality.

About two-thirds of hematopoietic stem cell transplants useautologous cells, generally extracted from peripheral blood by leukapheresis.The remainder are allogeneic transplants, most commonly using cells collecteddirectly from the bone marrow ofHLA-identical sibling donors (Figure 1 and Figure2).

The growth of hematopoietic stem cell transplantation has beenaccompanied by a coordinated, international effort to collect and analyze dataon transplant outcomes through the International Bone Marrow Transplant Registry(IBMTR), established in 1972, and the Autologous Blood and Marrow TransplantRegistry (ABMTR), established in 1990. Over 350 institutions in 47 countriescontribute data to the IBMTR, and over 250 institutions in North and SouthAmerica contribute data to the ABMTR. Participating centers submit data on theirconsecutive transplants to the IBMTR/ABMTR Statistical Center. The Centerreceives data on more than 12,000 new transplants each year and maintains adatabase that now includes information on more than 120,000 transplantrecipients.

Organizational Structureof the IBMTR/ABMTR

The activities of the IBMTR and ABMTR are supervised byAdvisory, Executive, and Working Committees and a joint IBMTR/ABMTR SteeringCommittee (Figure 3). The Advisory and Executive Committees review policies forthe use of IBMTR/ABMTR data and advise the scientific and statistical directorson administrative and scientific matters. The Working Committees design andconduct studies that are relevant to their subject area, consider proposals touse IBMTR/ABMTR data for specific studies, periodically assess and reviserelevant sections of IBMTR/ABMTR data collection forms, and plan and conductworkshops at IBMTR/ABMTR meetings. The Steering Committee establishes prioritiesfor scientific activities.

Since 1972, the IBMTR/ABMTR Statistical Center has been centralto Registry activities, coordinating data collection and management, andproviding statistical and administrative support for studies using Registrydata. The Statistical Center is an academic division of the Health PolicyInstitute of the Medical College of Wisconsin in Milwaukee.

Data Collection

The IBMTR/ABMTR collects data on two levels: registration andresearch. Registration data include disease type, age, sex, pretransplantdisease stage and response to chemotherapy, date of diagnosis, donor type, grafttype (bone marrow- and/or blood-derived stem cells), transplant regimen,posttransplant disease progression and survival, engraftment, graft-vs-hostdisease (GVHD), development of a new malignancy, and cause of death. AllIBMTR/ABMTR centers contribute registration data.

Research data are submitted on comprehensive report formscompleted for a subset of registered patients in IBMTR/ABMTR research centers.Research data include detailed pre- and posttransplant clinical information suchas disease subtype, tumor size and pathology, sites of disease, nontransplanttreatment of the primary disease, performance status, organ function, details ofthe transplant regimen including dose and schedule of high-dose therapy, graftmanipulation, supportive care, posttransplant toxicities, and functional status.

Both databases are longitudinal; patients are followed throughtheir transplant centers with yearly updates.

The Use of Multicenter Observational Databases

Transplant outcomes are influenced by many patient- anddisease-related factors (such as age, disease stage and prior treatment), aswell as transplant-related factors such as stem cell source, conditioningregimen, and prophylaxis for GVHD. Ideally, most transplant strategies would beevaluated by large randomized clinical trials. However, various factors limitthe application of randomized trials in hematopoietic stem cell transplantation.Many diseases treated with transplants are uncommon; thus, single centers maytreat only a few patients with a given disorder. This makes randomized trialsdifficult and also limits the ability to perform nonrandomized (phase II) trialswith sufficient power to detect meaningful effects. Small trials, even whenrandomized, may provide misleading results.[5]

New transplant technologies are rapidly being introduced, so theresults of prospective clinical trials may be obsolete before they arepublished. Some important transplant issues are not amenable to randomization,eg, differences in outcome associated with differences in donor type. In asystematic review of 255 transplant-related studies published between 1990 and1992, only 16 (6%) were randomized trials; most of these studies had fewer than100 patients.[6]

Even when randomized trials are performed, enrolled patients mayrepresent only a small proportion of the target population and may not berepresentative of the larger group.[7-10] The results of treatments administeredin these trials may differ from those obtained when the technology is morewidely applied. Most clinical trials focus on short- and intermediate-termoutcomes (1 to 5 years). However, there is a need for long-term follow-up oftransplant recipients because high-dose therapy may be associated with importanteffects, such as therapy-related cancers, that may not develop until years afterthe transplant was performed.

Observational databases may facilitate our understanding oftransplant outcomes by addressing questions that are difficult to address inrandomized trials. These include descriptions of transplant results in variousdisease states and patient groups; analysis of prognostic factors; evaluation ofnew transplant regimens; comparison of transplant with nontransplant therapy;defining intercenter variability in diagnosis, practice, and outcome; anddeveloping analytic approaches to evaluating transplantation outcomes and costs.

The value of observational studies in assessing treatmenteffects was highlighted by two recent articles published in the New EnglandJournal of Medicine.[11,12] Clinical databases may also be useful in developingoptimal designs for randomized studies and in interpreting the results of suchstudies.

Descriptive Studies

Databases that include a large number of centers involved inallogeneic and autologous transplantation are uniquely suited for descriptivestudies. This is particularly important in the case of rare diseases or morecommon conditions for which transplantation is infrequently performed. In thesesituations, single centers often have only one or a few cases, precludingmeaningful assessment of outcome. The publication of outcomes may be biased toresults that are particularly good or bad.[13] By combining data from manycenters and obtaining data systematically on all transplants, regardless ofoutcome, registries can provide a more precise and unbiased estimate of results.Examples of descriptive studies using IBMTR/ABMTR data are analyses oftransplants for Ph+ acute lymphoblastic leukemia,[14] Diamond-Blackfananemia,[15] chronic lymphocytic leukemia,[16] and paroxysmal nocturnalhemoglobinuria.[17]

Identification of Prognostic Factors

The heterogeneity and large numbers of patients reported to theIBMTR/ABMTR allow use of multivariate regression techniques to evaluateassociations between patient- and disease-related variables and outcome. Becausetransplant centers must report all consecutive transplant recipients, the fullrange of characteristics found in transplant patients is available for study.This is an important use of large observational databases, since many prognosticfactor studies are limited by small numbers, nonrepresentative populations,and/or insufficient detail on patient and disease characteristics.[18-21]

Examples of large prognostic factor studies using IBMTR/ABMTRdata include assessment of (1) risk factors for acute and chronic GVHD,[22,23]interstitial pneumonia,[24,25] and veno-occlusive disease of the liver[26]; (2)prognostic factors for relapse and leukemia-free survival after transplants foracute myelogenous leukemia, acute lymphoblastic leukemia, and chronicmyelogenous leukemia[27-30]; (3) risk factors for graft failure aftertransplants for severe aplastic anemia[31]; and (4) factors associated withthe outcome of autotransplants for metastatic breast cancer.[32]

Comparison ofTransplant Regimens

Ideally, new conditioning regimens, GVHD prophylaxis regimens,and other transplant maneuvers would be tested in large randomized trials.Randomization is the best available technique to minimize bias in treatmentassignment and to achieve balance in both known and unknown prognosticfactors.[33,34] An important limitation of randomized studies is the difficultyin accruing a sufficient number of patients to assess interventions withadequate statistical power. Registry databases contain valuable informationabout treatment efficacy for large numbers of patients. The major concern aboutdrawing inferences in nonrandomized settings is selection bias; treatments maybe assigned for reasons that are related to prognoses.[35,36] Observed benefitsmay therefore be related to prognosis rather than treatment.

It is possible, however, to make credible comparisons of variousstrategies using observational data, if there is sufficient clinical informationfor each patient to allow adjustment for the potentially confounding effects ofimportant prognostic variables.[11,12] The IBMTR has used this approach to(1) compare transplant regimens for aplastic anemia[37]; (2) compare GVHDprophylaxis regimens in patients with leukemia[38]; (3) compare the results ofHLA-identical sibling and HLA-mismatched related- and unrelated-donortransplants for leukemia[39]; (4) compare preparative regimens for allogeneictransplants in patients with acute lymphoblastic leukemia[40]; and(5) assess the role of T-cell depletion in allogeneictransplantation.[41,42]

Comparing TransplantationWith Alternative Treatments

With some diseases, transplantation is the only therapeuticoption; no other effective treatments are known. In other diseases (for example,acute leukemia), other potentially curative treatments exist. The question thenarises as to whether transplants are, in fact, superior to alternativetreatments. Rarity of the diseases treated, variable treatment philosophies, andlimited availability of donors, technologies, and resources have made evaluatingthese questions in randomized trials difficult.

It is also difficult to compare the published results oftransplant and nontransplant treatments directly because of differences inpatient selection as well as the inherent delay in performing transplants, whichleads to the truncation of early failures from most transplant series (ie,time-to-treatment bias).[43-46] Nevertheless, comparisons of transplantstrategies and alternative therapies can be made by combining observationaltransplant data with primary data compiled by groups studying nontransplantregimens, assuming that these datainclude adequate information on patient characteristics and similar criteria forassessing disease stage and outcome.[43-50]

The IBMTR/ABMTR has conducted several investigations comparingtransplant with nontransplant therapy in adults with acute myelogenous leukemiain first remission,[51] adults with acute myelogenous leukemia in secondremission,[52] adults with acute lymphoblastic leukemia in firstremission,[53-55] children with acute lymphoblastic leukemia in secondremission,[56] adults with chronic myelogenous leukemia in first chronicphase,[57] and women with metastatic breast cancer.[58] Chemotherapy data for these investigations have beencontributed by several cooperative oncology groups, including the EasternCooperative Oncology Group, the Pediatric Oncology Group, the German MulticenterAcute Leukemia Therapy Trials Groups, the German Acute Lymphoblastic LeukemiaTherapy Group, the Japan Acute Leukemia Study Group, the British MedicalResearch Council, and Cancer and Leukemia Group B.

Determining LateConsequences of Transplant

Large, longitudinal databases are particularly useful to studythe late effects of transplantation. The relative infrequency of latecomplications requires that large numbers of patients be followed for many yearsto determine precise estimates of risk and identify important prognosticfactors. The IBMTR/ABMTR has information on more than 15,000 persons whosurvived 3 or more years after transplantation.

The IBMTR/ABMTR, in collaboration with the National CancerInstitute and the Fred Hutchinson Cancer Center, recently completed a study ofcancers developing after allogeneic bone marrow transplants. They found thatsuch patients had an increased risk for brain, head and neck, liver and thyroidcancers, as well as melanoma and Hodgkin’s disease.[59-61] Another recentinvestigation evaluated late (ie, 2 or more years after transplant) causesof death in more than 6,000 allograft recipients, identifying disease recurrenceand GVHD as the most important causes of late treatment failure.[62] An analysisof late deaths after autotransplants is in progress, as is an analysis ofthe quality of life in long-term survivors of both allo- andautotransplantation.

Developing Statistical Methodology for Transplant Outcomes

Large databases are a rich resource for developing mathematicalmodels of posttransplant outcomes. Examples include addressing the problems ofadjusting for center effects following a proportional hazards regressionanalysis,[63] comparing transplant and nontransplant therapies,[43] comparingtherapies with diverging outcomes over time (eg, comparing autologous andallogeneic transplants where autologous transplants tend to have fewer earlydeaths but higher late failure from relapse),[64] modeling intermediate effectslike relapse and donor lymphocyte treatment,[65,66] and predicting the resultsof specific interventions.[67] IBMTR/ABMTR faculty statisticians activelyresearch optimal statistical methods for studying registry data and reformingother types of survival analyses.

Clinical Trials

Studies of observational data from multiple centers do notreplace the need for carefully conducted single-center phase I and II studies orsingle or multicenter randomized studies.[68] Rather, they offer a complementaryapproach. Large, representative databases can focus clinical trial effortson areas that are most likely to be productive and aid in clinical trialplanning, implementation, and interpretation. Precise estimates of outcomes andaccrual patterns can aid sample-size calculations and implementation plans.

Identification of appropriate centers for specific trials canincrease efficiency. Information on the most commonly used supportive caremeasures can increase the acceptability of protocols. Failure to accrue can besystematically addressed by studying the treatment of patients who are eligiblebut not enrolled. Comparison of clinical trial outcomes with observationaloutcomes can give insight about generalizability and patient selectionpractices.

Hematopoietic Stem Cell Transplantation: Indicationsand Outcomes

Use of hematopoietic stem cell transplantation grew dramaticallyover the past 30 years, and transplant strategies continue to change at a rapidpace. The remainder of this article uses IBMTR/ABMTR data to review the mostcommon indications for transplantation in North America, to describe outcomes oftransplantation for these disorders, and to identify the most common causes ofdeath after transplantation. These analyses used data on 11,347 patientstransplanted from 1995 through 1997 in 101 North American IBMTR/ABMTR researchcenters. These centers were able to provide complete follow-up through 1999 onover 85% of transplant recipients. The median number of transplants performed atthese centers during this period was 82 (range: < 10 to > 600). Autologoushematopoietic stem cells were used in 66% of these transplants, 20% were fromHLA-identical sibling donors, and 14% from other allogeneic donors. Thecharacteristics of the patient population are listed in Table1.

Changes in Indications

The indications for autologous and allogeneic transplantsdiffered. The majority of allogeneic transplants (74%) were performed forleukemia or preleukemia: 21% were for chronic myelogenous leukemia, 23% foracute myelogenous leukemia, 17% for acute lymphoblastic leukemia, 9% formyelodysplastic syndromes, and 4% for other leukemias. Fifteen percent wereperformed for other cancers, including non-Hodgkin’s lymphoma (9%), multiplemyeloma (4%), and Hodgkin’s disease (< 1%). The remainder were for aplasticanemia (4%), immune deficiencies (2%), inherited disorders of metabolism (2%),and other nonmalignant disorders.

Until recently, autotransplants were used solely to treatcancer. The most common indications for autotransplants in North America werebreast cancer (44%), non-Hodgkin’s lymphoma (24%), Hodgkin’s disease (8%),multiple myeloma (9%), acute myelogenous leukemia (4%), and a variety of othercancers. Two-year probabilities of survival after allogeneic and autologoustransplantation for the most common indications are shown in Table 2 and Table3. Themost striking recent change in the indications for autotransplantation over thepast decade was its use in breast cancer. In 1989, about 15% of autotransplantsin North America were for breast cancer, while in 1997, over 40% were for breastcancer. However, enthusiasm for the procedure decreased after the results (somepreliminary) of several studies were presented at the May 1999 American Societyof Clinical Oncology (ASCO) meetings.[69-71] These studies indicated little orno survival benefit associated with high-dose vs conventional-dose therapy forbreast cancer, although the need for further follow-up was emphasized by severalinvestigators. Preliminary ABMTR data suggest about a 40% to 50% reduction inthe number of autotransplants for breast cancer in 2000.

Earlier and Broader Use?

There is interest in using hematopoietic stem cell transplantsin several diseases for which transplants were not used or were only rarely usedin the past, and news of some promising results has emerged through anecdotalreports and phase II studies. These conditions include sickle-cell disease,inborn errors of metabolism, chronic lymphocytic leukemia, solid tumors such asovarian cancer and small-cell lung cancer, and autoimmune diseases such asmultiple sclerosis, systemic lupus erythematosus, and severe rheumatoidarthritis. Even taken all together, these currently account for fewer than 5% ofhematopoietic stem cell transplants. However, the prevalence of these diseasesis high and, if subsequent trials confirm the efficacy of transplants, thenumbers of persons treated with hematopoietic stem cell transplants couldincrease dramatically. In addition, hematopoietic stem cells are idealcandidates as vehicles for gene therapy, and their use in this capacity is beingexplored in several settings.

In the 1970s and early 1980s, most transplants were performed inpatients with refractory cancers, active infection, after receiving multipletransfusions, and with poor performance status. For example, in the 1970s, fewerthan 20% of transplants for leukemia were done in first remission or firstchronic phase. In the 1990s, about 60% of transplants for leukemia were done inpatients in first remission or first chronic phase. A trend toward earliertransplantation is seen particularly in chronic myelogenous leukemia; from 1984to 1985, the median interval between diagnosis and transplant was 17 vs 9 monthsfrom 1995 to 1997.

Many studies in diverse diseases demonstrate that earliertransplants are associated with lower risks for both transplant-relatedmortality and disease recurrence. However, many transplants (about half) areused as salvage therapy in advanced disease, particularly transplants forlymphoma and solid tumors and those using or unrelated donors.

Hematopoietic stem cell transplantation is now being performedin much older patients. In the 1970s, the median age of transplant recipientswas 17 years; from 1995 to 1997, the median age of allograft recipients was 31years and of autograft recipients 44 years. Although 40% of allografts and 68%of autografts performed between 1995 and 1997 were in patients older than 40years, only a minority (14% and 36%, respectively) were in patients older than50 years. This is important because the onset of the diseases for whichtransplants are most frequently performed is usually in older adulthood, oftenin the 50s or 60s.

The ability to use hematopoietic stem cell transplantation inolder patients makes it a useful treatment for many more patients and accounts,in part, for its increasing use as illustrated in Figure1. However, agelimitations are still an important impediment to the use of transplantation.Based solely on age considerations (upper limit of about 55 years for allogeneictransplantation and about 65 years for autotransplantation), only about 30% ofpatients with acute myelogenous leukemia, 30% of those with chronic myelogenousleukemia, 35% of those with non-Hodgkin’s lymphoma, and 15% of those withmultiple myeloma would be considered allograft candidates; correspondingpercentages for autotransplants would be 45%, 45%, 40%,and 30%.

Outcome of transplantation is affected by many patient- anddisease-related factors. In multiple studies, the most important of these weredisease type, disease stage, donor type, and recipient age. The 2-year survivalrates after autologous, HLA-identical sibling, and unrelated-donor transplants,by age and disease stage, for the most common transplant indications are shownin Table 2 and Table 3. Several groups reported posttransplant survival rates of 60%or greater in specific populations, generally younger patients with lessadvanced disease. Outcomes are poorer in patients with more advanced disease;however, these patients often have few alternative treatment options.

Causes of Death

An examination of the causes of death after transplantation canidentify the most common barriers to successful outcome. Among the 11,347patients described in Table 1, 5,761 (51%) died within 2 years of treatment. Theprimary causes of death, as reported by the transplantation centers, are shownin Figure 4. Among recipients of allogeneic transplants, GVHD accounted for 15%of deaths. Infection accounted for another 25% (most fatal infections were viralor fungal). Organ toxicity, both early and late, resulted in a significantnumber of deaths. Despite allogeneic graft-vs-tumor effects, recurrentmalignancy was a major contributor to death. Among autotransplant recipients,the majority of deaths were due to failure of high-dose therapy to control theprimary disease. Organ toxicity and infections were less important causes ofmortality, although they contributed to morbidity and cost.

Long-Term Survivors

Increasing use of hematopoietic stem cell transplantation andlower early mortality have led to large numbers of long-term transplantsurvivors. There are now more than 35,000 patients alive 5 or more years aftertransplantation. Although most survivors lead normal lives, transplantrecipients remain at risk for complications long after treatment. These includelate infections, chronic GVHD, cataracts, abnormalities of growth anddevelopment, thyroid disorders, chronic lung disease, and avascular necrosis ofbone. Therapy-related cancers are particularly common in children receivingcranial irradiation, in patients receiving high doses of total body irradiationas part of their pretransplant conditioning, and in those receivingautotransplants for lymphoma.[59-61,72-77] Some of this risk derives fromnontransplant treatment given for the primary disease. Additionally, risk ofrecurrence of the primary disease persists for many years, especially afterautotransplantation.

A recent IBMTR study suggests that although most 2-yearsurvivors are cured and have good performance status, excess mortality (ascompared to the general population) persists for 4 to > 10 years aftertransplantation, depending on the primary disease.[78] Similar data on relativemortality rates in autotransplant recipients are lacking. As reported to theIBMTR, among 7,075 persons surviving in remission 2 years afterallotransplantation for leukemia (1980-1994), 10-year survival was 83% (95%confidence interval, 81%-85%). About 40% of deaths were from recurrentleukemia, 15% from infection, 15% from GVHD, and 5% from second cancers.

As reported to the ABMTR, among 1,466 persons surviving inremission 2 years after autotransplantation for leukemia or lymphoma (1989-1994),10-year survival was 67% (range: 60%-74%). About 60% of deaths were fromrecurrent malignancy, 15% from second cancers, 10% from organ failure, and 5%from infection. Lifelong surveillance of transplant survivors is necessary as isincreased awareness of late complications among the many nontransplantphysicians who will care for these patients. Also important is a betterunderstanding of the impact of transplantation on the quality of life oftransplant survivors, an area currently under investigation by the IBMTR/ABMTRand others.

For a full list of committee members, ongoing Registry studies,and information on acquiring and using Registry data, please call theIBMTR/ABMTR Statistical Center (414-456-8325) or visit the Registry’s websiteat: www.ibmtr.org.

References:

1. Bach FH, Albertini RJ, Joo P, et al: Bone-marrowtransplantation in a patient with the Wiskott-Aldrich syndrome. Lancet2:1364-1366, 1968.

2. Gatti RA, Meuwissen HJ, Allen HD, et al: Immunologicalreconstitution of sex-linked immunological deficiency. Lancet 2:1366-1369, 1968.

3. Good RA, Meuwissen HF, Hong R, et al: Successful marrowtransplantation for correction of immunological deficit in lymphopenicagammaglobulinemia and treatment of immunologically induced pancytopenia. ExpHematol 19:4-10, 1969.

4. De Konig J, Dooren LJ, Van Bekkum DW, et al: Transplantationof bone-marrow cells and fetal thymus in an infant with lymphopenicimmunological deficiency. Lancet 1:1223-1227, 1969.

5. Fayers PM, Machin D: Sample size: How many patients arenecessary? Br J Cancer 72:1-9, 1995.

6. Niland JC, Gebhardt JA, Lee J, et al: Study design,statistical analyses, and results reporting in the bone marrow transplantationliterature. Biol Blood Marrow Transplant 1:47-53, 1995.

7. Begg CB: Selection of patients for clinical trials. SeminOncol 15:434-450, 1988.

8. Schmucker DL, Vesell ES: Underrepresentation of women inclinical drug trials. Clin Pharmacol Ther 54:11-15, 1993.

9. Gorkin L, Schron EB, Handshaw K, et al: Clinical trialenrollees vs nonenrollees: The Cardiac Arrhythmia Suppression Trial (CAST)Recruitment and Enrollment Assessment in Clinical Trials (REACT) project.Control Clin Trials 17:46-59, 1996.

10. Feinstein AR: An additional basic science for clinicalmedicine: II. The limitations of randomized trials. Ann Intern Med 99:544-550,1983.

11. Benson K, Hartz AJ: A comparison of observational studiesand randomized, controlled trials. N Engl J Med 342:1878-86, 2000.

12. Concato J, Shah N, Horwitz RI: Randomized, controlledtrials, observational studies, and the hierarchy of research designs. N Engl JMed 342:1887-1892, 2000.

13. Sutton AJ, Duval SJ, Tweedie RL, et al: Empirical assessmentof effect of publication bias on meta-analyses. Br Med J 320:1574-1577, 2000.

14. Barrett AJ, Horowitz MM, Ash RC, et al: Bone marrowtransplantation for Philadelphia chromosome-positive acute lymphoblasticleukemia. Blood 79:3067-3070, 1992.

15. Mugishima H, Gale RP, Rowlings PA, et al: Bone marrowtransplantation for Diamond-Blackfan anemia. Bone Marrow Transplant 15:55-58,1995.

16. Michallet M, Archimbaud E, Bandini G, et al: HLA-identicalsibling bone marrow transplantation in younger patients with chronic lymphocyticleukemia. Ann Intern Med 124:311-315, 1996.

17. Saso R, Marsh J, Evreska L, et al: Bone marrow transplantsfor paroxysmal nocturnal hemoglobinuria. Br J Haematol 104:392-396, 1999.

18. Harrell FE, Lee KL, Matchar DB, et al: Regression models forprognostic prediction: Advantages, problems, and suggested solutions. CancerTreat Rep 69:1071-1077, 1985.

19. Simon R, Altman DG: Statistical Aspects of prognostic factorstudies in oncology. Br J Cancer 69:979-985, 1994.

20. Greenland S: Power, sample size and smallest detectableeffect determination for multivariate studies. Stat Med 4:117-127, 1985.

21. Starmer CF, Lee KL: A data-based approach to assessingclinical interventions in the setting of chronic disease. Cancer Treat Rep66:1077-1082, 1982.

22. Gale RP, Bortin MM, van Bekkum DW, et al: Risk factors foracute graft-vs-host disease. Br J Haematol 67:397-406, 1988.

23. Atkinson K, Horowitz MM, Gale RP, et al: Risk factors forchronic graft-vs-host disease after HLA-identical sibling bone marrowtransplantation. Blood 75:2459-2464, 1990.

24. Weiner RS, Bortin MM, Gale RP, et al: Interstitialpneumonitis after bone marrow transplantation: Assessment of risk factors.Ann Intern Med 104:168-175, 1986.

25. Weiner RS, Horowitz MM, Gale RP, et al: Risk factors forinterstitial pneumonia following bone marrow transplantation for severe aplasticanemia. Br J Haematol 71:535-543, 1989.

26. Rozman C, Carreras E, Qian C, et al: Risk factors forhepatic veno-occlusive disease following HLA-identical sibling bone marrowtransplants for leukemia. Bone Marrow Transplant 17:75-80, 1996.

27. Gale RP, Horowitz M, Bortin M: IBMTR analysis of bone marrowtransplants in acute leukemia. Bone Marrow Transplant 4:83-85, 1990.

28. Gale RP, Horowitz MM, Weiner RS, et al: Impact ofcytogenetic abnormalities on outcome of bone marrow transplants in acutemyelogenous leukemia in first remission. Bone Marrow Transplant 16:203-208,1995.

29. Barrett AJ, Horowitz MM, Gale RP, et al: Marrowtransplantation for acute lymphoblastic leukemia: Factors affecting relapse andsurvival. Blood 74:862-871, 1989.

30. Goldman JM, Gale RP, Horowitz MM, et al: Bone marrowtransplantation for chronic myelogenous leukemia in chronicphase: Increased risk of relapse associated with T-cell depletion. AnnIntern Med 108:806-814, 1988.

31. Champlin RE, Horowitz MM, van Bekkum DW, et al: Graftfailure following bone marrow transplantation for severe aplastic anemia: Riskfactors and treatment results. Blood 73:606-613, 1989.

32. Rowlings PA, Williams SF, Antman KH, et al: Factorscorrelated with progression-free survival after high-dose therapy andhematopoietic stem cell transplantation for metastatic breast cancer. JAMA282:1335-1343, 1999.

33. Byar DP, Simon RM, Friedewald WT, et al: Randomized clinicaltrials: Perspectives on some recent ideas. N Engl J Med 295:74-80, 1976.

34. Green SB: Patient heterogeneity and the need for randomizedclinical trials. Control Clin Trials 3:189-198, 1982.

35. Garcoa C, Hidalgo M, Paz-Ares L, et al: Patient selection inhigh-dose chemotherapy trials: Relevance in high-risk breast cancer. J ClinOncol 15:3178-3184, 1997.

36. Rahman ZU, Frye DK, Bazdar AU, et al: Impact of selectionprocess on response rate and long-term survival of potential high-dosechemotherapy candidates treated with standard-dose doxorubicin-containingchemotherapy in patients with metastatic breast cancer. J Clin Oncol15:3171-3177, 1997.

37. Gluckman E, Horowitz MM, Champlin RE, et al: Bone marrowtransplantation for severe aplastic anemia: Influence of conditioning andgraft-vs-host disease prophylaxis regimens on outcome. Blood 79:269-275, 1992.

38. Ringdén O, Horowitz MM, Sondel P, et al: Methotrexate,cyclosporine, or both to prevent graft-vs-host disease after HLA-identicalsibling bone marrow transplants for early leukemia? Blood 81:1094-1101, 1993.

39. Szydlo R, Goldman JM, Klein JP, et al: Results of allogeneicbone marrow transplants for leukemia using donors other than HLA-identicalsiblings. J Clin Oncol 15:1767-1777, 1997.

40. Davies SM, Ramsay NKC, Klein JP, et al: Comparison ofpreparative regimens in transplants for children with acute lymphoblasticleukemia. J Clin Oncol 18:340-347, 2000.

41. Marmont AM, Horowitz MM, Gale RP, et al: T-cell depletion ofHLA-identical transplants in leukemia. Blood 78:2120-2130, 1991.

42. Champlin RE, Passweg JR, Zhang MJ, et al: T-cell depletionof bone marrow transplants for leukemia from donors other than HLA-identicalsiblings: Advantage of T-cell antibodies with narrow specificities. Blood95:3996-4003, 2000.

43. Klein JP, Zhang MJ: Statistical challenges in comparingchemotherapy and bone marrow transplantation as a treatment for leukemia, inJewel, Kimber, Lee, Whitmore (eds): Lifetime Data: Models in Reliability andSurvival Analysis, pp 175-186. Norwell, Massachusetts, Kulwer Academic Press,1996.

44. Begg CB, McGlave PB, Bennett JM, et al: A criticalcomparison of allogeneic bone marrow transplantation and conventionalchemotherapy as treatment for acute myelogenous leukemia. J Clin Oncol2:369-378, 1984.

45. Hermans J, Suciu S, Stijnen T, et al: Treatment of acutemyelogenous leukaemia: An EBMT-EORTC retrospective analysis of chemotherapy vsallogeneic or autologous bone marrow transplantation. Br J Ca Clin Oncol25:545-550, 1989.

46. Messerer D, Neiss A, Horowitz MM, et al: Comparison ofchemotherapy and bone marrow transplants using two independent clinicaldatabases. J Clin Epidemiol 47:1119-1126, 1994.

47. Moon TE, Jones SE, Bonadonna G, et al: Using a database ofprotocol studies to evaluate therapy: A breast cancer example. Stat Med3:333-339, 1984.

48. Davis K: The comprehensive cohort study: The use of registrydata to confirm and extend a randomized trial. Recent Results Cancer Res111:140-148, 1988.

49. Mantel N, Byar D: Evaluation of response-time data involvingtransient status: An illustration using heart-transplant data. J Am Stat Assoc69:81-86, 1974.

50. Turnbull BW, Brown BW, Hu M: Survivorship analysis of hearttransplant data. J Am Stat Assoc 69:74-80, 1974.

51. Gale RP, Büchner T, Zhang MJ, et al: HLA-identical siblingbone marrow transplants vs chemotherapy for acute myelogenous leukemia in firstremission. Leukemia 10:1687-1691, 1996.

52. Gale RP, Horowitz MM, Rees JKH, et al: Chemotherapy versustransplants for acute myelogenous leukemia in second remission. Leukemia10:13-19, 1996.

53. Horowitz MM, Messerer D, Hoelzer D, et al: Chemotherapycompared with bone marrow transplantation for adults with acute lymphoblasticleukemia in first remission. Ann Intern Med 115:13-18, 1991.

54. Zhang MJ, Hoelzer D, Horowitz MM, et al: Long-term follow-upof adults with acute lymphoblastic leukemia in first remission treated withchemotherapy or bone marrow transplantation. Ann Intern Med 123:428-431, 1995.

55. Oh H, Gale RP, Zhang MJ, et al: Chemotherapy vsHLA-identical sibling transplants for adults with acute lymphoblastic leukemiain first remission. Bone Marrow Transplant 22:253-257, 1998.

56. Barrett AJ, Horowitz MM, Pollock BH, et al: Bone marrowtransplants from HLA-identical siblings as compared with chemotherapy forchildren with acute lymphoblastic leukemia in a second remission. N Engl J Med331:1253-1258, 1994.

57. Gale RP, Hehlmann R, Zhang MJ, et al: Survival with bonemarrow transplantation vs hydroxyurea or interferon for chronic myelogenousleukemia. Blood 91:1810-1819, 1998

58. Berry DA, Broadwater G, Klein JP, et al: High-dose vsstandard chemotherapy in metastatic breast cancer: Comparison of Cancer andLeukemia Group B trials with data from the Autologous Blood and MarrowTransplant Registry (submitted, J Clin Oncol).

59. Curtis RE, Rowlings PA, Deeg HJ, et al: Solid cancers afterbone marrow transplantation. N Engl J Med 336:897-904, 1997.

60. Rowlings PA, Curtis RE, Passweg JR, et al: Increasedincidence of Hodgkin disease following allogeneic bone marrow transplant. J ClinOncol 17:3122-3127, 1999.

61. Socié G, Curtis RE, Deeg HJ, et al: New malignant diseasesafter allogeneic marrow transplantation for childhood acute leukemia. J ClinOncol 18:348-357, 2000.

62. Socié G, Veum-Stone J, Wingard JR, et al: Long-termsurvival and late deaths after allogeneic bone marrow transplantation. N Engl JMed 341:14-21, 1999.

63. Anderson PK, Klein JP, Zhang MJ: Testing for center effectsin multicenter survival studies: A Monte Carlo comparison of fixed and randomeffects. Stat Med 18:1489-1500, 1999.

64. Klein JP, Zhang M: Confidence regions for the equality oftwo survival curves under the proportional hazards model (in press). LifetimeData Anal 2001.

65. Klein JP, Keiding N, Shu Y, et al: Summary curves forpatients transplanted for chronic myeloid leukemia salvaged by a donorlymphocyte infusion: the current leukemia free survival curve. Br J Haematol109:148-152, 2000.

66. Klein JP, Szydlo RA, Craddock C, et al: Estimation ofcurrent leukemia-free survival following donor lymphocyte infusion therapy forpatients with leukemia who relapse after allografting: application of amultistate model. Stat Med 19(21):3005-3016, 2000.

67. Keiding N, Klein JP, Horowitz MM: Multistate models andoutcome prediction in bone marrow transplantation (in press). Stat Med 2001.

68. Pocock SJ, Elbourne DR: Randomized trials or observationaltribulations? N Engl J Med 342:1907-1909, 2000.

69. Stadtmauer EA, O’Neill A, Goldstein LJ, et al: Phase IIIrandomized trial of high-dose chemotherapy (HDC) and stem cell support (SCT)shows no difference in overall survival or severe toxicity compared tomaintenance chemotherapy with cyclophosphamide, methotrexate, and fluorouracil(CMF) for women with metastatic breast cancer who are responding to conventionalinduction chemotherapy: The "Philadelphia" Intergroup Study (BPT-1). NEngl J Med, 342:1069-1076, 2000.

70. Rodenhuis S, Bontenbal M, Beex L, et al: Randomized phase IIstudy of high-dose cyclophosphamide, thiotepa, and carboplatin in operablebreast cancer with 4 or more axillary lymph nodes (abstract). Proc Am Soc ClinOncol 19:74a, 2000.

71. Peters W, Rosner G, Vredenburgh J, et al: A prospective,randomized comparison of two doses of combination alkylating agents asconsolidation after CAF in high-risk primary breast cancer involving ten or moreaxillary lymph nodes: Preliminary results of CALGB 9082/SWOG 9114/NCIC MA-13(abstract). Proc Am Soc Clin Oncol 18:1a, 1999.

72. Witherspoon RP, Fisher LD, Schoch G, et al: Secondarycancers after bone marrow transplantation for leukemia and aplastic anemia. NEngl J Med 321:784-789, 1989.

73. Witherspoon RP, Deeg HJ, Storb R: Secondary malignanciesafter marrow transplantation for leukemia or aplastic anemia. Transplant Sci4:33-41, 1994.

74. Miller JS, Arthur DC, Litz CE, et al: Myelodysplasticsyndrome after autologous bone marrow transplantation: An additional latecomplication of curative cancer therapy. Blood 83:3780-3786, 1994.

75. Darrington DL, Vose JM, Anderson JR, et al: Incidence andcharacterization of secondary myelodysplastic syndrome and acute myelogenousleukemia following high-dose chemoradiotherapy and autologous stem-celltransplantation for lymphoid malignancies. J Clin Oncol 12:2527-2534, 1994.

76. Traweek ST, Slovak ML, Nademanee P, et al: Clonal karyotypichematopoietic cell abnormalities occurring after autologous bone marrowtransplantation for Hodgkin’s disease and non-Hodgkin’s lymphoma. Blood84:957-963, 1994.

77. Deeg HJ, Socié G, Schoch G, et al: Malignancies aftermarrow transplantation for aplastic anemia and Fanconi anemia: A joint Seattleand Paris analysis of results in 700 patients. Blood 87:386-392, 1996.

78. Socié G, Veum-Stone J, Wingard JR, et al: Long-termsurvival and late deaths after allogeneic bone marrow transplantation. N Engl JMed 341:14-21, 1999.