|Home | About | Journals | Submit | Contact Us | Français|
Diminished bone mineral density (BMD) is a known complication after bone marrow transplantation (BMT) in childhood 1–3. Endocrine dysfunction, irradiation, corticosteroids and chronic graft versus host disease (cGVHD) have been identified as risk factors for diminished BMD in childhood cancer survivor populations 4–9. To date a comprehensive understanding of the mechanisms and risks for diminished BMD after childhood BMT has not be described.
Measurement of BMD density in childhood cancer survivors has also been problematic. In children the World Health Organization (WHO) criteria are used for the clinical diagnosis of osteopenia/osteoporosis based on Z-scores calculated from age; sex; and race; matched reference data. Many childhood BMT survivors are small for age, making the standard Z-score comparisons based on age/sex alone inaccurate, systematically underestimating BMD in smaller children 10. Current recommendations suggest three dimensional (volumetric) measures of BMD should be used in this population 11.
Dual energy x-ray absorptiometry (DEXA) is the most frequently used technique to assess BMD; it is readily available and associated with very low exposure to ionizing radiation. DEXA acquires BMD measures in two dimensions, with calculation necessary to obtain volumetric measures 12. We identified no studies that compared the areal and calculated volumetric BMD measures from DEXA scans in childhood BMT survivors.
Maintenance of appropriate BMD is multi-factorial including the influence of other components of body composition, fat mass and fat-free mass. The associations between BMD and fat mass and fat-free mass vary according to the population studied but may explain up to 31% of the variance in BMD in healthy populations 13. We identified no studies that examined the associations between BMD and body composition in childhood BMT survivors. Knowledge of these associations may be important for understanding the BMD outcomes and may guide treatment decisions
The aims of this study were to describe the lumbar spine BMD of BMT survivors; compare the DEXA scan acquired areal and volumetric measures of BMD; explore the relationship between BMD measures and body composition; and identify disease and treatment factors that predict diminished BMD
Childhood BMT survivors who were at least two years post transplant, were recruited from an established survivor clinic and IRB-approved recruitment letters were mailed to those whose names and addresses were generated from the institutional BMT database. Twelve survivors were recruited from 160 mailings; the remaining 34 were recruited from clinic visits. This protocol was approved by the Johns Hopkins Review Board and informed consent was obtained for all participants. Parental consent and patient assent was obtained for those younger than 18 years.
Medical record review was used to gather data on age at time of BMT; years since BMT; autologous versus allogeneic; preparative regimen including total body irradiation (TBI); central nervous system (CNS) irradiation (not including TBI only); and history of chronic graft-versus-host-disease (cGVHD). In addition, medical records were reviewed to determine prednisone equivalent cumulative mg/m2 steroid dose, including pre-transplant regimens until resolution of cGVHD. Patient/family report and medical record reviews were used to assess endocrine function which was categorized as 1) Normal= no endocrinopathies or thyroid stimulating hormone suppression therapy only; and, 2) Abnormal = Growth hormone and/or gonadal hormone replacement. Physical exam was used to determine pubertal development by Tanner stage (pubic hair). Height was measured to the nearest centimeter with a stadiometer. Kilograms of fat-free and fat mass were obtained from DEXA data and were converted to fat mass index (FMI) and fat-free mass index (FFMI) as follows: (kg)/(Height (m))2.
Lumbar spine (L1–L4) BMD was determined using DEXA scans, providing areal BMD (g/cm2) and age/sex/ethnic matched Z-scores derived from United States Food and Drug Administration approved healthy controls (GE, Prodigy). Volumetric BMD (g/cm3) values for L1–L4 were calculated to determine apparent BMD (volumetric BMD = BMD (4/(π × width))14. Volumetric BMD Z-scores were calculated from published lumbar spine volumetric BMD (g/cm3) reference data on 444 age/sex matched healthy controls obtained using the same DEXA equipment and calculated with the formula described above 15. The Z-Scores from each measure were used to identify survivors meeting WHO definitions of osteopenia (Z = −1 to −2.5) and osteoporosis (Z= < − 2.5) 16.
STATA statistical software version 10 was used for analysis 17. Z- scores for areal (g/cm2) and volumetric (g/cm3) measures were calculated as described above and compared to WHO criteria for osteopenia and osteoporosis. Pearson’s correlation was used to examine the relationships between areal and volumetric BMD measures.
Univariate and multivariate analysis for volumetric z-scores and predictor variables were performed. Multivariate model selection was performed with forward selection. Covariates considered for the regression model include; age, age at time of BMT, years out from BMT, Tanner stage, endocrine function, CNS irradiation, total body irradiation, type of BMT, cGVHD, cumulative steroid dose, FMI and FFMI 4–6. An alpha level of 0.05 was used to determine significance.
Subject characteristics are listed in Table 1. The distribution of transplant indications and types were reasonably representative of distributions for the pediatric BMT program at Johns Hopkins. In addition, approximately 25% of BMTs at Johns Hopkins are done for non-malignant diseases. In this study only 8 survivors had non-malignant conditions including aplastic anemia and X-linked autoimmune/allergic disregulation syndrome, but no survivors had Thalassemia. Three of the survivors who received CNS irradiation for a hematologic malignancy also had a preparative regimen including TBI. Tanner stages were bimodal and were appropriately reflective of age.
BMD outcomes for BMT survivors were analyzed using Z-scores for areal and volumetric measures. Utilizing the z-scores for the areal lumbar BMD, 22% (10/46) of survivors met criteria for osteopenia and 2% (1/46) for osteoporosis. Utilizing volumetric lumbar spine BMD Z-scores, 22% (10/46) of survivors met criteria for osteopenia and none met criteria for osteoporosis. Characteristics of survivors with diminished volumetric BMD are listed in Table 3.
Comparison of the areal and volumetric measures included analysis of values for survivors who were small for age and differences in WHO diagnoses for the two measures. As expected, areal lumbar and volumetric lumbar BMD Z-scores were correlated (Pearson, r=0.73, p= <0.001). For the 8 survivors who were below the 3rd percentile for height (according to Center for Disease Control, United States growth chart), 7 had higher z-scores with the volumetric measures. The only survivor meeting criteria for osteoporosis had an areal Z-score of −3.7 and volumetric Z-score of -2.25. This survivor was an 18-year-old male with a height of 153cm (well below the 3rd percentile).
Differences in classification of osteopenia/osteoporosis between the two measures were identified (Figure 1). Thirty-six percent (4/10) of survivors with osteopenia using areal measures had normal volumetric Z-scores, and the median height Z-scores calculated from National Health and Nutrition Examination Survey (NHANES) data 18 for this group was −1.37 (iqr 1.27). Thirty percent (3/10) of survivors with osteopenia using volumetric measures had normal areal z-scores, and the median height Z-score for this group was 0.39 (iqr 1.5).
As supported by our data, volumetric measures are the recommended method for evaluating BMD in childhood cancer survivors and therefore used for the predictive models. On univariate analysis only age (CI: −0.127–−.022, p = 0.007) and CNS irradiation (CI: −1.870–−0.089, p= 0.032) had statistically significant associations with volumetric z-scores. Multiple linear regression was performed with volumetric z-scores as the response variable (Table 2). Controlling for cumulative steroid dose; CNS irradiation, female gender and older age at the time of testing were significantly associated with lower volumetric Z-scores. FMI was positively associated with volumetric Z-scores.
This study identifies new findings relevant to the detection of and risk factors for diminished BMD in survivors of BMT during childhood. First, volumetric measures appear to be the most appropriate way to evaluate survivors who are small for age. Second, CNS irradiation, older age at time of testing, and female sex are risk factors for diminished BMD. Third, we found a positive association between FMI and BMD that has not previously been described in childhood BMT survivors.
The appropriate measurement of BMD in BMT survivors is an important topic still under debate. While the need to use volumetric measures of BMD has been recommended the most suitable exam has not been identified. The most frequently used radiologic exams used to measure volumetric BMD are the quantitative computed tomography (QCT) and DEXA scans. While the QCT directly measures volumetric BMD, it is associated with higher doses of ionizing radiation and is therefore utilized less to measure BMD. DEXA scans are recognized as the “gold standard” for measuring BMD with minimal ionizing radiation but require manual calculation of volumetric BMD10. Identification of the safest and most accurate approach to measuring BMD after BMT is important to both clinical practice and future research.
In this study the prevalence of diminished BMD (osteopenia/osteoporosis) was similar whether areal or volumetric measures of the lumbar spine were used (areal 11/46, volumetric 10/46). Although as expected areal and volumetric lumbar spine BMD measures were significantly correlated, there were important differences in classifications of diminished BMD for individual survivors. These incongruent findings are significant for individual survivors since the interventions for, and monitoring of diminished BMD is based on Z-score findings. Using thresholds defined by the WHO for osteopenia/osteoporosis, 30% of patients with osteopenia using volumetric Z-scores had normal areal Z-scores, and were survivors of average height. Conversely, 36% of patients with osteopenia using areal z-scores had normal volumetric z-scores; these survivors were small for age and the higher volumetric score likely reflects correction of underestimation of the areal measurement. We observed that 88% of survivors who were below the 3rd percentile for height in this study had higher BMD Z-scores with volumetric measures, further evidence that DEXA-derived volumetric measures may correct for underestimation of BMD in this population.
The findings from our study suggest that DEXA-derived volumetric calculation of lumbar BMD is the most appropriate method for BMT survivors who are small for age. The most appropriate measure to use for diagnosis and subsequent treatment of osteopenia/osteoporosis in survivors who are not small of age is not determined by the results of this study, but we recommend standardizing a volumetric measurement for all childhood BMT survivors.
In addition to measurement issues, the understanding of mechanisms necessary to develop and maintain optimal BMD in cancer survivors is evolving. The importance of understanding the full spectrum of factors that contribute to BMD is critical in evaluating causes and developing interventions for diminished BMD in BMT survivors. Body composition (FM and FFM) is associated with BMD in healthy populations but to date the evaluation of these associations in childhood BMT survivors is missing in the literature 13.
In this study we found that FMI was associated with higher volumetric z-scores. Fat mass has been associated with increased BMD in obese children and adolescents and is thought to result from the increased mechanical loading placed on the bones 19. Another possible explanation is elevated leptin levels associated with obesity. Leptin stimulates chondrocytes and is hypothesized to contribute to BMD in obese individuals 19;20.
The positive effects of FM on BMD would be accompanied by health risks associated with obesity such as diabetes and heart disease. Therefore it would be inappropriate to suggest increasing FM beyond healthy recommendations in survivors when other means of maintaining and increasing BMD exists. Studies in other populations have shown that exercise training that increases lean mass and fitness can maintain or increase BMD despite the fact that exercise also decreased body fat 21. Results of the present study would suggest that exercise interventions designed to increase BMD in childhood BMT survivors should recognize the important relationships between bone, muscle and fat and capitalize on interventions that will have a positive impact on overall health.
Other treatment and demographic associations were identified in our analysis. Our findings confirm previous studies that found CNS irradiation was associated with decreased BMD8;22;23. In our study several explanations exist for this finding. In 75% of the survivors with CNS irradiation the lumbar spine was included in the field, thus a direct effect of radiation decreasing chondrocytes may have contributed to this association with lower lumbar BMD 24. The doses and sites of CNS irradiation varied and the sample was too small to examine site and dose associations but would be important to consider in future studies.
In other CNS-irradiated patient populations, antiepileptics and physical disability have been associated with diminished BMD and may help explain the relationship identified in this study 25. No survivors in our study were currently being treated with antiepileptics but some may have had previous exposure to these agents. Finally, two of our survivors who received CNS irradiation were treated for CNS tumors which require higher doses of radiation in addition to surgical interventions. Both of these survivors have significant ataxia which limits the types of physical activity they participate in, especially high impact activities which are known to be important for improving and maintaining BMD.
In addition, CNS irradiation may be associated with pituitary dysfunction, resulting in thyroid, gonadal and/or growth hormone deficiency, all of which can negatively affect BMD. While all of the survivors with diagnosed hormone deficiency were receiving hormone replacement therapy, there may have been more subtle, untreated abnormalities in the CNS irradiated group. For example, subnormal androgen levels have been identified in female BMT survivors and testosterone replacement improves BMD in women who have hypogonadal functioning due to hypopituitarism 26. Endocrine function and replacement data were determined by medical record review and patient report for our study, but levels were not measured. Future studies should consider further evaluation of endocrine function and the impact that current replacement strategies have on BMD.
We also found older age at the time of testing was associated with lower volumetric Z-scores. The lower volumetric Z-scores in older survivors emphasizes the need for screening and intervention in childhood survivors so optimal peak bone mass can be achieved and maintained.
Female BMT survivors were also found to have lower volumetric z-scores. In healthy older females, osteoporosis and osteopenia are major health concerns; compromised BMD after BMT will make this an even greater concern for female survivors. Diminished BMD in female childhood BMT survivors has previously been reported and attributed to gonadal dysfunction 1. As described above, endocrine abnormalities not readily diagnosed in female survivors, including androgen deficiency, may help to account for the negative association between female sex and volumetric BMD in our study. Seventy five percent (6/8) of the females with abnormal endocrine function in our study were on gonadal HRT at the time of testing. Research is needed to identify the optimal HRT after BMT in female survivors.
Treatment with steroids has been implicated in diminished BMD prior to and after childhood BMT 27, 7. Cumulative steroid doses, which included therapies prior to BMT through the resolution of cGVHD were not associated with lower volumetric z-scores in either univariate (β = 0.000; p=0.150) or multivariate analysis in our study (β=0.000; p=0.073). Possible explanations include that other studies were conducted in childhood BMT survivors closer to the time of treatment (up to 12 months post BMT)7. Evidence exists that the effects of steroids on BMD are reversible and that BMD increases after discontinuing steroid therapy28. In our study the mean years out from BMT was 8 (2–22) and no patients were actively taking steroids, these findings suggest that diminished BMD associated with steroid treatment is time limited and may recover over time in childhood BMT survivors.
Treatment of diminished BMD after childhood cancer has not been well studied. Most studies have had small sample sizes, included only hematologic malignancies and utilized bisphosphonates 29, 30, 31, 32. Biphosphonate therapy has also been used in the treatment and prevention of diminished BMD in adults undergoing BMT 33, 34. While these studies have shown that bisphosphonates are well tolerated and improve BMD in these populations more information is necessary to make treatment recommendations. Certainly treatment of any underlying conditions such as endocrinopathies are recommended in survivors with diminished BMD. Treatment with calcium, vitamin D and exercise have been studied in healthy pediatric populations with varying degrees of effectiveness 35, 36, 37, 38. An ongoing, longitudinal study of these interventions in childhood ALL survivors may validate their effectiveness in this population39.
Maintenance of appropriate BMD is multifactorial and the etiology of diminished BMD may change over time in childhood BMT survivors. Our study confirms that childhood BMT survivors are at risk for abnormal BMD, primarily osteopenia. Osteopenia in younger populations has been shown to predict future osteoporosis and fractures 40. Early detection of diminished BMD will allow for appropriate interventions and may prevent future disability.
We recommend the use of volumetric measures of BMD in survivors who are small for their age. Further research comparing other methods, including quantitative CT, is needed to confirm that DEXA-derived volumetric measures are most appropriate for all BMT survivors.
FMI was found to have a positive association with BMD. Interventions that increase FFM and improve overall fitness may help to maintain BMD in survivors who are obese. The significance of fat mass is an important addition to the understanding of BMD outcomes for BMT survivors and may guide interventions in individual patients. Additional risk factors for diminished BMD identified in this study include treatment with CNS irradiation and female sex. Further research is necessary to elucidate the underlying mechanisms for these associations and should include the evaluation of possible endocrine etiologies and interventions.
Individual National Research Service Award (NRSA), National Institute of Nursing Research (NR010038-01)
American Cancer Society, Doctoral Nursing Scholarship, 112191
General Clinical Research Center (GCRC), Johns Hopkins Bayview Medical Center
Dedication: Victoria Mock, PhD, RN, FAAN
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.