Blood Journal
Leading the way in experimental and clinical research in hematology

Late effects in survivors of chronic myeloid leukemia treated with hematopoietic cell transplantation: results from the Bone Marrow Transplant Survivor Study

  1. K. Scott Baker,
  2. James G. Gurney,
  3. Kirsten K. Ness,
  4. Ravi Bhatia,
  5. Stephen J. Forman,
  6. Liton Francisco,
  7. Philip B. McGlave,
  8. Leslie L. Robison,
  9. David S. Snyder,
  10. Daniel J. Weisdorf, and
  11. Smita Bhatia
  1. From the University of Minnesota Cancer Center, Blood and Marrow Transplant Program, and the Department of Pediatrics and the Department of Internal Medicine, University of Minnesota, Minneapolis, MN; and Hematology and Hematology/Bone Marrow Transplant, City of Hope Cancer Center, Duarte, CA.


The purpose of this study was to analyze medical late effects among patients with chronic myeloid leukemia (CML) treated with hematopoietic cell transplantation (HCT). Subjects included 248 CML survivors who received an HC transplant (related donors [RDs], n = 150; unrelated donors [URDs], n = 70; or autologous, n = 28) and had survived at least 2 years, and a comparison group of 317 siblings. Subjects completed a 238-item survey on medical late effects. Compared with siblings, survivors were at a higher risk of developing ocular, oral health, endocrine, gastrointestinal, musculoskeletal, neurosensory, and neuromotor impairments. Multivariate analysis limited to RD and URD recipients found that chronic graft-versus-host disease (cGVHD) was associated with a higher risk of hypothyroidism, osteoporosis, cardiopulmonary, neurosensory, and neuromotor impairments. Overall health was reported as excellent, very good, or good in 78% of subjects, although those with cGVHD were more likely to report poor overall health. URD survivors were more likely to report a need for assistance with routine activities and that their current health prevented work or school attendance. This study demonstrates that HCT survivors, regardless of donor type, have a high prevalence of long-term health-related complications. However, adverse medical late effects with significant morbidity were uncommon. Chronic GVHD is the most important predictor of adverse medical late effects and poor overall health.


Hematopoietic stem cell transplantation is a potentially curative therapy for patients with chronic myeloid leukemia (CML), and before the imatinib (Gleevec; Norvartis, Basel, Switzerland) era, was considered to be the treatment of choice.1 Three-year survival rates of 85% can be achieved in good-risk patients (those younger than 40 years of age, in first chronic phase, and receiving a transplant less than 1 year from diagnosis) who undergo matched sibling donor (MSD) transplantation.2 Similar encouraging results can be achieved in good-risk patients who receive matched unrelated donor (URD) transplants, in whom disease-free survival (DFS) rates of 60% to 70% have been reported.3-7 Although older age at diagnosis, advanced phase of disease (accelerated or blast crisis), and transplantations performed after the first year from diagnosis are associated with poorer results, cures can still can be achieved in a substantial proportion of patients.3,4,6,8 However, deaths from late relapses and chronic graft-versus-host disease (cGVHD) continue to occur many years after transplantation.9,10

Few data exist on the quality of the long-term survival in patients who have undergone hematopoietic stem cell transplantation (HCT) for CML, and few studies have specifically addressed issues related to medical complications in long-term survivors of CML treated with HCT. Data from the few clinical studies available are hindered by limited sample sizes, the inclusion of multiple diagnoses in the study cohorts, and lack of appropriate control groups. In a small study of patients with CML (n = 26), late complications such as osteoporosis, cataracts, hypothyroidism, and second cancers were reported in 7% to 18% of patients who survived for 10 or more years.9 This study also found lower levels of physical functioning and general health compared with the general population, although all subjects still reported overall satisfaction with HCT.11 In a larger study that compared long-term outcome for 316 CML patients who received cyclophosphamide (CY) and busulfan or total body irradiation (TBI), more than 90% of the patients had normal health status or minimal impairment, and between 72% and 88% had returned to school or work.12 However, the study evaluated only a very limited spectrum of medical complications. Of note, cataracts developed in 47% of patients, and avascular osteonecrosis developed in 10%, both of which were associated with exposure to TBI and cGVHD. Late pulmonary disease was reported in 15% of patients with CML regardless of preparative regimen.

The purpose of the current study was to analyze and describe the frequency of, and predictors for, the development of a broad spectrum of medical complications, functional limitations, and overall general health in long-term survivors of CML treated with HCT. Additionally, the results from this study would be useful to clinicians faced with the task of determining the long-term risks and benefits of HCT therapy for CML patients, and would provide data to form the basis for the long-term follow-up and evaluation of these survivors.

Patients and methods

This analysis is part of the Bone Marrow Survivor Study (BMT-SS), a collaborative study between City of Hope Cancer Center (Duarte, CA) and the University of Minnesota (Minneapolis, MN) that was established in 2000 to evaluate outcomes in a cohort of individuals with a diagnosis of cancer or other life-threatening illness and who were treated with HCT. Eligibility criteria include the following: (1) HCT for hematologic and nonhematologic malignancies and other life-threatening disorders; (2) date of transplantation between January 1, 1974, and December 31, 1998; and (3) survival of 2 years or longer after HCT regardless of disease status. This analysis is limited to English-speaking CML survivors who received either an autologous or an allogeneic HC transplant from a related donor (RD) or a URD, and who were alive at the time of the initial interview. For comparison purposes, a group of siblings was enrolled into the study by assembling a random sample of study participants stratified on the basis of diagnosis, age, sex, ethnic background; asking these participants to recruit their siblings into the study; and then asking the siblings to complete the same questionnaire as the participants. The informed consent process, and the study protocols and documents, were approved by the Human Subjects Research Review Committees at the City of Hope Cancer Center and the University of Minnesota. Each eligible participant, or his or her parent or legal guardian for participants younger than 18 years of age at the time of interview, provided informed consent for the study.

Data collection

Participants completed the BMT-SS questionnaire, a 238-item survey assessing medical late effects, current medical conditions, medication use, health status, health behaviors, pregnancy history, demographic characteristics, socioeconomic indicators, insurance coverage, and other information. The BMT-SS questionnaire asks participants to report impairments of organ systems, limitations that interfere with daily function, and the impact of these impairments and/or functional limitations on daily life either at home, at school, or at work. The questionnaire has a yes/no/don't know format for the majority of the questions and, for yes responses, requires the participant to indicate the year the impairment was first diagnosed by a doctor or health care provider. Some questions use a Likert or ordinal-type response choice to score degree of impairment or dysfunction. This instrument was developed for use by the Childhood Cancer Survivor Study13 and was subsequently modified to address topics specifically related to the HCT population in the current study. The BMT-SS questionnaire has been validated on a random sample of 100 HCT survivors. The agreement with medical records was excellent (percentage agreement adjusted for chance, κ> 0.8) for musculoskeletal, cardiovascular, pulmonary, and endocrine impairments and for graft-versus-host disease, and moderate (κ, 0.4-0.7) for second cancers, central nervous system disorders, and eye problems.14 Treatment information was obtained from the HCT databases at each institution.


A total of 381 eligible patients were identified. Of these, 21 patients (6%) were lost to follow-up; 100 (26%) declined participation; and 12 (3%) had yet to complete the study questionnaire and are considered passive refusals. A total of 248 survivors completed the questionnaire, which represents a 65% response rate among those potentially eligible, and 69% among those successfully contacted. Study participants did not differ statistically from eligible nonparticipants across categories of sex (P = .16), time since transplantation (P = .59), age at transplantation (P = .12), type of transplant (P = .23), chronic GVHD (P = .59), or conditioning regimen (P = .56).

At the time of this analysis, a total of 317 siblings had been recruited into the BMT-SS. The comparative evaluations for this analysis were not limited to siblings of CML patients, but included siblings of all patients who were enrolled in the study and had completed a study questionnaire. Characteristics of the sibling participants were similar to the transplant recipients, except that a higher proportion of siblings were white and female (Table 1).

View this table:
Table 1.

Characteristics of study participants

Data analysis

The prevalence of organ or organ system impairments was calculated by tabulating the affirmative responses to specific questions presented in the BMT-SS questionnaire, if the participant indicated that the organ system complication occurred after the date of HCT. Participants were divided into 3 transplantation groups on the basis of donor source: (1) autologous, (2) related donor, and (3) unrelated donor. All participants had myeloablative chemotherapy, with or without TBI, in preparation for their transplants.

Means, medians, standard deviations, and ranges were calculated for age at diagnosis, age at interview, and time since transplantation. Frequencies and percentages were calculated for transplant type, sex, age at diagnosis, age at interview, time since transplantation, each medical late effect, each organ system impairment, each functional limitation, and inability to participate in work, school, or home activities. Chi-square statistics or Fisher exact tests were used to evaluate differences in variable characteristics between patients and siblings, and between transplant types among patients. Cumulative incidence was calculated by means of a competing risk analysis based on the Markov Chain Approach of Aalen and Johansen.15

The presence or absence of cGVHD was represented as a dichotomous variable and was determined by the participant's response to the question: “Have you ever been told by a doctor or other health care professional that you have, or have had chronic graft versus host disease?” Owing to the small number of survivors in this study who had received autologous transplants and the few patients who did not receive TBI, the multivariate analyses were limited to those patients who had received an allogeneic HC transplant and TBI in the preparative regimen. For each of the dichotomous outcome variables of organ system impairment, functional limitation, or inability to participate in work, school, or home activities, logistic regression analyses were conducted to simultaneously account for the effects of donor type (RD versus URD), sex, and the presence of cGVHD after adjusting for age at transplantation and time since transplantation. Age at transplantation was entered into the regression analyses as a continuous variable, and time since transplantation as a 2-level class variable of 2 to 5 years and 6 years or longer. There was no evidence for statistically significant interactions between age at time of HCT, or time since HCT, and other variables in the models. Interaction terms for donor type and sex, and for donor type and cGVHD, were also evaluated. Confounding by age, sex, time since transplantation, and cGVHD were examined by looking at the strength and the precision of the estimate in both full and reduced models. The phi coefficient for the correlation between cGVHD and donor type was calculated prior to entering both terms in the model to evaluate the assumption of independence. Correction for multiple comparisons was not performed because these are primarily descriptive data that were not testing a specific primary hypothesis. SAS version 8.2 was used for all analyses (SAS Institute, Cary, NC).



Characteristics of the study participants are shown in Table 1. Of the 248 individuals who participated in this study, 28 (11%) received an autologous HC transplant; 150 (61%) received their HC transplant from a related donor; and 70 (28%) received it from an unrelated donor. Slightly more than half of the subjects were male (55%), and 81% were white. Of the study participants, 95% were age 21 years or older at transplantation and 50% were older than age 45 years when interviewed (median age at interview, 45 years; range, 21-69 years). Median time between transplantation and study participation was 7 years (range, 2 to 27 years), and approximately two thirds (63%) of the study population had survived 6 years or longer following transplantation. The majority of subjects in this study (94%) received TBI as part of their transplantation-preparative regimen. There were no significant differences between the allogeneic donor groups (RD versus URD) in the distribution of sex, race, age at transplantation, age at study participation, time since transplantation, or treatment institution.

Organ and organ system impairments

Comparison with siblings. Characteristics of the sibling participants were similar to the transplant recipients, except that a higher proportion of siblings were white and female. Table 2 shows the prevalence and odds ratios (ORs) (adjusted for age at interview and sex) for selected medical late effects comparing all patients who received an HC transplant with the sibling group. Note that while the odds ratios for several of the outcomes evaluated are significantly elevated, the prevalence of many of them is still quite low. Compared with siblings, transplantation survivors were 7.5 times more likely to experience sequelae related to the ocular system, including being 15.3 times more likely to develop cataracts (95% confidence interval [CI], 8.2-28.2), and 3.9 times more likely to have dry eyes (95% CI, 2.4-6.4). Oral health impairments were 2.5 times more likely in survivors compared with siblings and included dry mouth (OR, 16.8; 95% CI, 5.2-54.6) and difficulty chewing or swallowing (OR, 7.8; 95% CI, 2.6-24). The relative odds of having an endocrine impairment was 3.2 times (95% CI, 2.1-5.0) higher in survivors compared with siblings and consisted primarily of a 2.7 times (95% CI, 1.6-4.5) higher risk of developing hypothyroidism and a 4.9 times (95% CI, 2.3-10.5) higher risk of diabetes. Survivors were also more likely than siblings to have musculoskeletal abnormalities, including being 6.4 times (95% CI, 2.9-14) more likely to develop osteoporosis and 12 times (95% CI, 1.5-93.7) more likely to have had avascular necrosis. HCT survivors were no more likely than siblings to report significant cardiopulmonary impairments such as arrhythmias, congestive heart failure, myocardial infarction, coronary heart disease, hypertension, or stroke. Survivors were, however, 4.2 times more likely to report problems with exercise-induced shortness of breath compared with the sibling group. Other cardiopulmonary impairments reported infrequently in the transplantation group were angina (n = 3; 1.2%), pericarditis (n = 4; 1.6%), heart valve disorders (n = 2; 0.8%), blood clots (n = 10; 4%), and lung fibrosis (n = 0), and there was no difference between HCT survivors and siblings in the prevalence of these events. The relative odds of gastrointestinal abnormalities were higher in survivors than in siblings (OR, 2.2; 95% CI, 1.3-3.6), including a 2.4 times (95% CI, 1.2-4.3) higher risk of gallstones and a 2.9 times (95% CI, 1.0-7.8) higher risk of hepatitis. Neurosensory impairments were 2.8 times more frequent in survivors than in siblings, with the most significant risks being the loss of, or abnormal sense of, taste or smell (OR, 35.2; 95% CI, 8.6-144) and loss of, or abnormal sense of, touch (OR, 2.7; 95% CI, 1.6-4.4). Neuromotor impairments were 4.4 times more likely to occur in survivors compared with siblings, attributable primarily to a 5.1 times (95% CI, 2.8-9.3) higher risk of problems with balance, tremor, or weakness.

View this table:
Table 2.

Frequency and percentage of self-reported medical late effects comparing transplant recipients with siblings, including adjusted odds ratios

To further examine the impact of chronic GVHD, an analysis was performed stratifying the comparison with siblings on the basis of the presence or absence of chronic GVHD (Table 3). For the most part, these comparisons revealed the same associations shown in Table 2, which compared the entire transplantation group with the siblings; the exceptions were as follows: participants without chronic GVHD were not significantly different from siblings in the frequency of oral health impairments (18.3% versus 12.9%; P = .13); problems chewing or swallowing (4.6% versus 1.3%; P = .08); hypothyroidism (10.5% versus 7.3%; P = .15); avascular necrosis (2% versus 1%; P = .32); cardiopulmonary impairments overall (26% versus 25.9%; P = .99); exercise-induced shortness of breath (5.9% versus 2.5%; P = .10); and gastrointestinal impairments (15% versus 9.1%; P = .07). This indicates that the development of these impairments, in particular, occurs more frequently in association with the presence of chronic GHVD but that even participants without chronic GVHD experience a higher frequency of many late effects, such as cataracts, dry mouth, diabetes, osteoporosis, and neurosensory and neuromotor impairments, than sibling controls (Table 3).

View this table:
Table 3.

Frequency and percentage of self-reported medical late effects comparing siblings with patients by chronic graft-versus-host disease status (limited to most frequent outcomes associated with chronic GVHD)

Comparison by donor source. To determine if stem cell source was associated with the development of medical late effects after HCT, comparisons were made between transplant recipients by donor source (Table 4). The reported prevalence of ocular involvement (primarily cataracts and dry eyes) was significantly higher among recipients of RD or URD transplants compared with those who received autologous transplants (P = .01). The cumulative incidence of cataracts was 53% (95% CI, 12%-93%) at 15 years for autologous HCT survivors; 63% (95% CI, 50%-77%) for RD survivors; and 67% (95% CI, 51%-83%) for URD survivors (Figure 1). The prevalence of oral impairments, endocrine abnormalities, musculoskeletal abnormalities, or gastrointestinal dysfunction did not differ significantly by donor type. The cumulative incidence of hypothyroidism at 10 years after HCT was 13% (95% CI, 0%-32%) for survivors of autologous HCT; 8% (95% CI, 3%-12%) in the RD group; and 19% (95% CI, 3%-33%) in the URD group (Figure 1). The overall frequency of having any cardiac impairment was significantly higher in the URD group than in either the RD or the autologous transplant group (45.7% versus 29.3% versus 17.9% respectively; P < .05). However, when individual cardiopulmonary outcomes were examined, the only outcome with a significant difference was a higher reported prevalence of exercise-induced shortness of breath in the URD recipients (17.1%), compared with the RD (6.7%) and autologous (3.6%) HC transplant recipients (P < .05). The overall prevalence of neurosensory or neuromotor impairments did not differ among groups, although URD transplant recipients were significantly more likely to report either a loss of, or an abnormal sense of, taste or smell compared with autologous transplant recipients. Past or current cGVHD was reported more frequently by recipients of URD HC transplants compared with RD HC transplant recipients (59% versus 36%; P < .05), although the distribution of sites of involvement were not different in the groups.

View this table:
Table 4.

Frequency and percentage of self-reported medical late effects comparing transplant types

Figure 1.

Cumulative incidence of cataracts, hypothyroidism, and second malignancies.

Subsequent malignancies

Table 4 also shows the prevalence of subsequent malignancies, with 21 survivors reporting a second malignancy, and 4 reporting a third malignancy. Histologic types of second primary malignancies included basal cell carcinoma of the skin (n = 12), skin melanomas (n = 3), sarcomas (n = 2), lymphoma (n = 2), Bowen disease (n = 1), and a malignant mole. The third malignancies included lymphoma (n = 1) and basal cell carcinoma (n = 3). The cumulative incidence of a subsequent malignancy at 10 years after HCT was 7% (95% CI, 0%-17%) for survivors after autologous HCT; 5% (95% CI, 1%-9%) after RD HCT; and 15% (95% CI, 2%-27%) after URD HCT. The 20-year follow-up for the RD and URD groups showed that the cumulative incidence of a subsequent malignancy increased to 16% (95% CI, 15%-31%) for RD and 15% (95% CI, 3%-27%) for the URD group.

Reproductive outcomes

Seven men and 3 women reported the presence of 14 pregnancies following HCT (autologous HCT, n = 1; URD HCT, n = 1; RD HCT, n = 12), resulting in 7 live births, 1 stillbirth, 3 miscarriages, and 1 medical abortion; the outcome of 2 pregnancies was not reported. Five of 7 males and 2 of 3 females who reported post-HCT pregnancies also reported fertility problems.

Functional limitations

Figures 2 and 3 compare the prevalence of functional limitations, again comparing survivors with siblings and by donor type.

Figure 2.

Proportion of functional limitations in a comparison of all transplant recipients with siblings.

Figure 3.

Proportion of functional limitations in a comparison of transplant types. * indicates significant difference from other proportions for that outcome.

Comparison with siblings. As shown in Figure 2, HCT survivors were more likely than siblings to report the need for assistance with routine activities such as housework or shopping (12.6% versus 2.5%; P < .001), and that their current health prevented them from attending school or work (26.3% versus 2.2%; P < .001). Siblings were more likely than HCT survivors to rate their health as excellent, very good, or good (94.3% versus 78.2%; P < .001), and HCT survivors more likely to rate their health as fair or poor (21.8% versus 5.7%; P < .001).

Comparison by donor type. As shown in Figure 3, participants in the URD HC transplant recipient group were more likely than those with an RD or an autologous transplant to report the need for assistance with routine activities such as housework and shopping (21.7% versus 8.0% versus 14.3% respectively; P < .05). Both the URD (40.6%) and the autologous HC transplant recipients (35.7%) were more likely to report that their current health prevented school or work attendance, compared with the RD group (18%; P < .05). More than 70% of all subjects reported their overall health as excellent, very good, or good, although URD HC transplant recipients were more likely to report poor or fair health (30.0%) than were the RD (18%) or autologous (21.4%) HC transplant recipients (P < .05).

Predictors of medical late effects, functional limitations, and poor overall health

Table 5 shows the multivariate models quantifying the relative odds of developing hypothyroidism; osteoporosis; cardiopulmonary impairment; an impaired sense of touch; or balance impairment, tremor, or weakness. Also shown in Table 5 are the relative odds of reporting the need for assistance with personal care or routine activities, of reporting that health interfered with school or work attendance, and of reporting poor overall health. The correlation between cGVHD and donor type was small (ϕ= 0.24), indicating that the 2 variables were largely independent of one another; thus, both variables were included in the models.16 There was no evidence of multiplicative interaction between donor type and sex, or between donor type and age at transplantation, for any of these impairments; thus, no interaction terms were included in the models. Overall, the strongest risk factor for development of these selected medical late effects was the presence of cGVHD. Female sex was identified as a risk factor for osteoporosis. After donor type, sex, age at transplantation, and time since transplantation were accounted for, the relative odds of developing hypothyroidism were 3.21 times (95% CI, 1.4-7.5) higher in those with cGVHD when compared with those without cGVHD. The relative odds of developing osteoporosis were 2.65 times (95% CI, 1.0-7.2) higher; cardiopulmonary impairments, 1.74 times (95% CI, 0.9-3.2) higher; impaired sense of touch, 2.11 times (95% CI, 1.1-4.2) higher; and balance impairment, tremor, or weakness, 2.06 times (95% CI, 1.0-4.1) higher in those who reported cGVHD when compared with those who did not report cGVHD. The presence of cGVHD was also a predictor of self-reported poor health; individuals who reported the presence of cGVHD were 2.41 times (95% CI, 1.2-4.9) more likely to report poor overall health than those without cGVHD after donor type, sex, time since transplantation, and age at transplantation were accounted for. No risk factors were identified for fertility problems in either males or females after adjustment for age and time since transplantation. Self-reported functional limitations were associated with female sex and URD HCT, but not cGVHD. Females were 3.40 times (95% CI, 1.3-8.9) more likely than males to report the need for assistance with routine activities such as shopping and housework after adjustment for donor type, cGVHD, time since transplantation, and age at transplantation. Persons receiving a URD HC transplant were 2.64 times (95% CI, 1.0-6.7) more likely to report the need for assistance with routine activities and 2.43 times (95% CI, 1.2-4.9) more likely to report that their current health prevents work or school attendance than those with a related-donor transplant.

View this table:
Table 5.

Relative odds of having a selected medical late effect, a fertility problem, or a functional limitation, or reporting poor health based on donor type, presence or absence of chronic graft-versus-host disease, and sex


This large 2-center study of long-term CML survivors evaluated a wide spectrum of potential late effects after HCT. The study had the added benefit of comparison of the long-term health of the HCT survivors with that of a random sample of siblings of transplant recipients. Importantly, nearly 80% of CML HCT survivors report the overall rating of their health as excellent, very good, or good, which is an indication that HCT is a treatment option that gives patients long-term benefit. However, in comparison with the siblings in the study, CML survivors after HCT were found to have a wide range of medical complications occurring at a higher-than-expected frequency. The majority of these conditions, however, are not necessarily associated with significant morbidity. Of interest was a lack of increased risk for any serious cardiopulmonary impairment (with the exception of exercise-induced shortness of breath), although 50% of subjects were still younger than age 45 years at time of interview. In general, it is apparent that the medical impairments evaluated in this study do have an impact on restriction in daily life, since approximately 12% of transplant recipients reported the need for assistance with routine activities and about one quarter reported having impairments in their current health that prevented them from attending school or work. Similar issues were reported by Kiss et al11 in a small number of long-term CML HCT survivors (n = 26) whose physical functioning and general health were rated lower than population norms and who reported less satisfaction with their physical strength.

Unrelated-donor transplantation is known to be associated with a higher prevalence of acute morbidity and mortality when compared with transplantation from other donor sources, but chronic effects in long-term survivors after unrelated donor HCT have not been well studied. Owing to the increasing use of unrelated donors for CML and other diseases, one of the goals of the current study was to examine the impact of donor source on the occurrence of medical late effects. Our study included only a small number of autologous transplants, so comparisons with that group should be considered preliminary. The findings presented here suggest that long-term survivors of CML after allogeneic HCT using a cyclophosphamide and TBI–based preparative regimen have a similar prevalence of medical late effects, irrespective of the donor source. In the multivariate analysis controlling for cGVHD and other factors, the only adverse outcomes found to be associated with the use of URD were limitations in function and limitations in role performance, including the need for assistance with routine activities and current health preventing work or school attendance. This corresponds with a higher proportion of survivors after URD transplant who rated their health as fair or poor as compared with RD or autologous transplant recipients. For URD transplant recipients, these results are similar to those reported by Baker et al,17 where a significant correlation was found between role retention and quality of life in a cohort of 135 bone marrow transplantation survivors. From our analysis, it is not clear why URD recipients have a higher frequency of functional limitations than RD recipients, since it does not appear to be entirely associated with a higher frequency of medical late effects. Chronic GVHD was the best predictor of medical late effects in the cohort of allogeneic HCT survivors. Thus, the primary risk in terms of late effects associated with the use of URD stem cells is in relation to the higher prevalence of chronic GVHD in these recipients.

The results of this study must be evaluated in the context of several limitations. The data were collected by self-report, and thus some degree of misclassification certainly exists. However, validation studies using medical records as a gold standard showed that self-report was good, at least for major organ impairments.14 The study questionnaire asks the participant to report conditions that were diagnosed by a physician or other health care provider; thus, subjects who infrequently seek medical care, or do not have post-HCT problems such as chronic GVHD that require ongoing follow-up, may have undiagnosed conditions that were not reported. Relapse after HCT may also affect the health status of survivors, although for this study subjects surviving 2 years or longer after HCT were eligible regardless of disease status. The study questionnaire was not structured to ascertain disease status at the time of study participation; however, institutional records indicated that the majority of subjects were in remission at time of last contact. In addition, we have not collected data on potentially eligible subjects who survived 2 years but then subsequently died from relapse, complications related to therapy or chronic GVHD, or second cancers. More comprehensive analyses of second cancers within the entire HCT cohorts (including survivors and deceased patients who developed a new malignancy at any time after HCT) from the University of Minnesota18 and City of Hope19 have been published previously. Another limitation is that we have not fully evaluated the impact of time since transplantation on the prevalence of many late effects. Some outcomes are likely to be more common early after HCT, and others may increase in prevalence with increasing length of follow-up. Time since transplantation was accounted for in our multivariate models, but a separate analysis would be required to carefully examine the impact of time since transplantation on the development of many of these long-term outcomes. Finally, a participation rate of 65% could potentially introduce some bias if the prevalence of outcomes among the nonrespondents differed systematically from that of the study participants. However, our data indicate that there were no significant differences between the respondents and the nonrespondents when the clinical and demographic characteristics of the 2 groups were compared.

A final issue to consider is the relevance of our findings in the current era of treatment for CML. The recent availability of the tyrosine kinase inhibitor imatinib which produces major cytogenetic responses in more than 85% of CML patients,20 now provides adefinite alternative to HCT. In chronic-phase patients, imatinib is reported to result in complete cytogenetic response rates of 76% and major cytogenetic response rates of 87%,20 and responses are seen even in patients with accelerated phase or blast crisis.21,22 This has led to a significant impact on the number of transplantations being performed for CML. However, at this point the durability of response to imatinib is unknown, and the development of resistance to imatinib both in vitro and in vivo has been documented.23-26 The decision regarding which patients should proceed to HCT, and when, has become difficult. Whether treatment with imatinib and potential delay in the timing of HCT will potentially alter HCT outcomes, including late effects, remains to be seen. Providing a clear understanding of the long-term adverse effects in HCT survivors will help physicians counsel patients regarding the choice between potentially curative therapy with HCT and the continued use of imatinib or other agents.

In summary, this comprehensive study presents new data from a large cohort of long-term CML survivors treated with HCT, provides comparisons with a healthy sibling control group, and evaluates risk factors for the development of late effects after HCT. Our data would suggest that the majority of CML survivors after HCT fare quite well, and that a large proportion of the long-term complications in this cohort are not potentially life threatening and are amenable to medical intervention. These findings are relevant to a large number of patients with CML who are surviving after HCT and provide guidance for the spectrum of late effects that should be monitored long term in these survivors.


  • Reprints:

    K. Scott Baker, Pediatric Blood and Marrow Transplant Program, University of Minnesota, 420 Delaware St SE, Mayo Bldg, Room D557, Mayo Mail Code 484, Minneapolis, MN 55455; e-mail: baker084{at}
  • Prepublished online as Blood First Edition Paper, June 1, 2004; DOI 10.1182/blood-2004-03-1010.

  • Supported by National Institutes of Health (NIH) CA78938, CA30206 and K23-CA85503; and the Leukemia & Lymphoma Society Scholar Award.

  • The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 U.S.C. section 1734.

  • Submitted March 18, 2004.
  • Accepted April 29, 2004.


View Abstract