Bone health in children and adolescents after allogeneic stem cell transplantation

High prevalence of vertebral compression fractures

Abstract

BACKGROUND

This cross-sectional study evaluated the overall bone health and the prevalence of vertebral complications after stem cell transplantation (SCT) in prepubertal children and adolescents.

METHODS

A total of 44 children and adolescents (median age, 10 years) were evaluated at a median of 3.8 years after SCT for areal bone mineral density (aBMD) with dual-energy X-ray absoptiometry and for vertebral fractures with instant vertebral assessment. Pretransplant and posttransplant medications and nutritional parameters were recorded, and plasma levels of vitamin D, calcium, phosphate, and parathormone were measured.

RESULTS

Of the 44 patients, 16 (36%) had a BMD Z-score of <−1.0. The patients with low BMD did not differ from the others with regard to their clinical or biochemical characteristics. Prepubertal patients had better BMD Z-scores at all sites compared with pubertal or postpubertal subjects. This was evident especially at the hip, in which the median aBMD Z-score in prepubertal patients (−0.2; range, −0.5 to +1.7) was found to be significantly higher than in pubertal (−1.1; range, −1.5 to +0.4) and postpubertal (−1.1; range, −2.6 to +0.5) patients (P = .03). Five patients (11%) had a history of peripheral fractures. Nine patients (20%) had vertebral compression fractures, which were asymptomatic in 7 patients.

CONCLUSIONS

Approximately one-third of patients who had undergone allogeneic SCT in childhood were found to have a reduced BMD before reaching adulthood. This was due in part to inadequate BMD gain during the pubertal years. The high prevalence of asymptomatic vertebral compression fractures calls for the systematic assessment of spinal health during the posttransplantation follow-up. Cancer 2007. © 2007 American Cancer Society.

Patients with a chronic illness are at risk of developing symptomatic osteoporosis due to factors related to the underlying condition, medications, nutritional problems, immobilization, or secondary hormonal disturbances.1 The long-term survivors of solid organ2, 3 or hematopoietic stem cell transplantation (SCT)4–8 have been shown to have an imbalance between bone formation and resorption and decreased bone mass and, consequently, an increased susceptibility to fractures.9 In adult recipients of SCT, a decrease of 5% to 9% in the lumbar spine or femoral neck bone mineral density (BMD) occurs during the first 6 months after transplantation.4 Recovery may take place later, especially in the lumbar spine,10 but the femoral neck BMD may deteriorate even further during the long-term follow-up.6, 10

The risk factors associated with impaired bone health in children after SCT have been reported to be delayed puberty and short stature,7, 11 prepubertal status at SCT, female gender, and hypogonadism.7 However, to our knowledge, data regarding the clinical significance of reduced BMD (ie, fractures) in pediatric SCT survivors are sparse. In addition, little is known concerning BMD in prepubertal children after SCT.

In the current study, we evaluated various parameters of bone health in 44 children and adolescents to elucidate the prevalence and clinical significance of skeletal adverse effects of childhood SCT.

MATERIALS AND METHODS

Study Population

Fifty-nine consecutive SCT patients who were reporting for regular follow-up visits to the Pediatric Hematology-Oncology outpatient clinic at the Helsinki University Central Hospital between April 2004 and December 2005 were invited to participate in this cross-sectional study. The inclusion criteria were 1) allogeneic SCT performed at least 6 months prior to the study entry and 2) age >4.0 years. Forty-four patients (75%) were willing to participate. The patients not willing to participate were 3 girls and 12 boys who were similar to the study subjects with respect to the underlying diagnosis, age at SCT, the median follow-up after SCT, and the type of donor, but were older (median age, 13.7 years; range, 7.7–19.5 years) at the time of their follow-up visit than the patients in the study cohort (median age, 10.1 years; range, 4.7–22.6 years). The study protocol was approved by the Institutional Review Board. Informed consent was obtained from age-appropriate study participants or their parents/guardians.

Clinical Characteristics

The previous medical data, including fractures, posttransplantation late endocrine effects, ongoing endocrine replacement therapies, maximum grade of acute graft-versus-host disease (GVHD), and activity of chronic GVHD,12 were collected from the hospital records. The duration of glucocorticoid (pretransplantation and posttransplantation combined) and cyclosporine therapies were recorded. The total cumulative dose of glucocorticoids, including both the pretransplantation and posttransplantation periods, was calculated and converted to prednisone equivalents. The cumulative dose of methotrexate was calculated. The irradiation doses and fields were recorded. A 3-day dietary recall was obtained from each patient; the patients were asked to record trade names and quantities of all consumed food items and beverages as well as calcium and vitamin D supplements. Based on these recordings, the average daily intake of calcium and vitamin D were calculated for each patient using a software program based on Finnish dietary references (AIVO 2000-Diet32 [version 1.4.2.1]). Patients or their parents/guardians completed a questionnaire regarding the child's physical activity and fracture history. The questionnaire regarding physical activity is unvalidated but has been specifically developed for studies assessing bone health in chronically ill children13 and gives an estimate of the overall physical activity of the child. Physical activity was graded as follows: none indicated exercise less than once a week, moderate indicated exercise once weekly, and good indicated exercise at least twice weekly. Age-appropriate physical activity at school, in daycare, or during leisure time was considered as exercise. The patients were clinically assessed and their height, weight, and pubertal maturation were recorded according to the method of Tanner.14. Height was transformed into standard deviation (SD) units (Z-scores) by comparing them with the Finnish growth curves.15 Weight was expressed as the height-adjusted value according to the Finnish standards.15

Imaging Studies

A plain X-ray of the left hand was obtained and bone age was determined according to the method of Greulich and Pyle.16 Areal BMD (aBMD) for the lumbar spine (L1–L4), total hip, and whole body was measured with dual energy X-ray absorptiometry (DXA) (Discovery A; Hologic, Inc, Waltham, MA). The aBMDs (lumbar spine, total hip, and whole body) were transformed into Z-scores by comparing the measured values with age-specific and sex-specific reference data for the equipment, as provided by the manufacturer. In addition, the Z-scores were corrected for skeletal maturation by using bone age instead of chronologic age. This correction, rather than correction for “height age,” was regarded as appropriate based on anticipation of delayed bone age in several patients. In the absence of a Finnish national database, the BMD reference data for white children in the U.S., which were provided by the manufacturer, were used.

Anteroposterior and lateral images of the thoracic and lumbar spine (instant vertebral assessment [IVA]) were obtained with the DXA scanner and evaluated for the presence of vertebral compression fractures; a height reduction of 20% in the anterior, middle, and/or posterior vertebral height was considered significant and height reductions of <20% were regarded as normal.17 The vertebral changes were classified as normal, mild, or severe anterior wedge deformities or mild or severe compression deformities according to the classification by Makitie et al.17 IVA images were used rather than spinal radiographs to avoid the significant radiation exposure caused by radiography. If compression fractures were suspected, they were confirmed by spinal radiographs if possible. If compression deformities were present in the lumbar spine, the lumbar aBMD values of the patient were omitted from the analyses.

Laboratory Tests

Plasma calcium, phosphate, and alkaline phophatase were analyzed using a photometric method. Serum 25-hydroxyvitamin D [25(OH)D] was assessed by liquid chromatography and plasma parathormone was assessed by an immunoluminometric method. Serum concentrations of insulin-like growth factor-1 (IGF-1) (assessed by radioimmunoassay [RIA]; DiaSorin, Stillwater, Minn) and IGF-1-binding protein-3 (IGFBP-3) (assessed by competitive RIA; Nichols Institute Diagnostics, San Clemente, Calif) were analyzed for all patients, and serum luteinizing hormone (assessed by immunofluorometric method), follicle-stimulating hormone (assessed by immunofluorometric method), and testosterone (assessed by liquid chromatography-mass spectrometry)/estradiol (assessed by RIA) concentrations were analyzed for children aged >8 years.

Statistical Analysis

SPSS software (SPSS Inc, Chigaco, Ill) was used in the statistical analyses. The parameters were presented as medians with corresponding ranges. The nonparametric Kruskal-Wallis test was used in comparing patients with acute lymphoblastic leukemia (ALL) with the other diagnosis groups. The Mann-Whitney U test was used to compare the clinical and biochemical parameters between patients with and those without reduced BMD and vertebral fractures.

RESULTS

Patient Characteristics

The study included 44 patients (27 males [61%]) ranging in age from 4.7 to 22.6 years (median age, 10.1 years) (Table 1). The median age at SCT was 6.3 years (range, 1.0–18.6 years) and the median follow-up after transplantation was 3.8 years (range, 0.6–9.2 years). The underlying diagnoses were acute myeloid leukemia (AML) (11 patients), ALL or T-cell lymphoma (18 patients), aplastic anemia (4 patients), chronic myeloid leukemia (CML) (3 patients), myelodysplasia (2 patients), histiocytosis (2 patients), chronic granulomatous disease (2 patients), Diamond-Blackfan anemia (1 patient), and Wiscott-Aldrich syndrome (1 patient). The conditioning regimen was comprised of total body irradiation (TBI) of 6 grays (Gy) (1 patient), 10 Gy (35 patients), or 12 Gy (4 patients) with high-dose cytarabine, cyclophosphamide, or melphalan. One patient received total lymph node irradiation (6 Gy) and 3 patients were primed without TBI. The donors were human leukocyte antigen (HLA)-identical siblings (20 patients) or unrelated donors (24 patients) from the Finnish National Bone Marrow Donor Registry or international registries. Two patients received cord blood stem cell grafts and the other patients received bone marrow grafts. Cyclosporine combined with short-term methotrexate was used as prophylaxis against GVHD. Due to differences in the pretransplantation therapy of the underlying disease, ALL patients had received a larger cumulative dose of glucocorticoids and methotrexate compared with patients with AML, CML, myelodysplasia, or those patients with a nonmalignant disease.

Table 1. Characteristics of the SCT Patients (Shown as The Medians [Range])

	All patients (N = 44)	ALL (N = 18)	AML/CML/ Premalignant* (N = 16)	Nonmalignant† (N = 10)	P
SCT indicates stem cell transplantation; ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CML, chronic myeloid leukemia; F, female; M, male; SD, standard deviation; BMI, body mass index; TBI, total body irradiation; MUD, matched unrelated donor; GVHD, graft-versus-host disease; Cya, cyclosporine; GH, growth hormone. The Kruskal-Wallis test was used to compared differences between the patients with ALL and the other groups. * Premalignant, no chemotherapy: myelodysplastic syndrome. † Nonmalignant: severe aplastic anemia, Diamond-Blackfan anemia, Wiscott-Aldrich syndrome, Langerhans cell histiocytosis, familial hemophagocytic histiocytosis, chronic granulomatous disease. ‡ Fractures in the extremities after diagnosis of the underlying disease.
Gender (F/M)	17/27	6/12	8/8	3/7	.90
Age at study, y	10.1 (4.7–22.6)	11.6 (5.6–20.5)	11.0 (4.7–22.6)	7. 6 (6.2–15.6)	.14
Age at SCT, y	6.3 (1.0–18.6)	7.3 (1.4–18.6)	4.8 (1.3–15.9)	5.3 (1.0–10.3)	.22
Height (SD)	−0.5 (−4.1–2.5)	−0.5 (−3.7–1.8)	−0.4 (−4.1–2.5)	−0.6 (−3.4–1.7)	.92
BMI (kg/m^2)	16.8 (13.2–27.6)	17.3 (13.2–27.6)	16.2 (13.4–23.2)	16.6 (13.8–18.7)	.13
TBI (n)	41	18	16	7	.14
Donor, sibling/MUD	19/25	7/11	8/8	4/6	.92
GVHD (n)
Acute grade 2–4	20	10	5	5	.42
Chronic (extensive)	11 (8)	5 (3)	4 (4)	2 (1)	.52
Prednisone
Cumulative dose, mg/kg	242 (0–2175)	235 (120–1070)	13 (0–763)	262 (0–2175)	.004
Cumulative duration, mo	4 (0–24)	5 (2–24)	0.1 (0–18)	9 (0–24)	.003
Cya duration, mo	6 (0–27)	7 (0–18)	6 (0–27)	8.5 (1–21)	0.93
Endocrinology (n)
Hypothyroidism	1	1	–	–
GH deficiency	3	2	1	–	.89
Adrenocortical failure	1	1	–	–
Hypogonadism	7	3	3	1	.81
Fractures‡	5	2	1	1	.70

The median height Z-score was −0.5 (range, −4.1 to +2.5); 5 patients had a height Z-score of <−2.0. The median height-adjusted weight was +3% (range, −26 to +57%) and 5 patients were overweight (height-adjusted weight of >+20%). Eight girls and 18 boys were prepubertal, 10 patients (4 of whom were girls) were pubertal (Tanner stages 2–3), and 8 patients (5 of whom were girls) were postpubertal (Tanner stages 4–5). The patients with hypothyroidism (1 patient), adrenocortical failure (1 patient), and growth hormone insufficiency (3 patients) were receiving appropriate substitution therapy. Two patients were receiving estrogen replacement therapy due to hypergonadotropic hypogonadism. Two female patients had been receiving estrogen therapy but had normal gonadotropin levels and normal menstrual cycles at 9 months and 2 years, respectively, after withdrawal of the substitution therapy. During the study, 3 additional patients (2 girls and 1 boy) were found to have supranormal gonadotropin levels suggestive of gonadal insufficiency.

Five patients (11%) had a history of peripheral fractures after diagnosis of the underlying disease (a total of 8 fractures, 1–2 fractures per patient). All fractures (3 in the tibia, 3 in the antebrachium, and 2 in the phalanges) had resulted from a low-impact trauma.

Twenty-four of the 44 patients (55%) completed a questionnaire concerning physical activity and 26 patients (59%) returned the dietary recall recording. Physical activity was moderate or good in 18 of /the 24 patients (75%). The median daily yields of calcium and vitamin D were 172% (range, 58–319%) and 103% (range, 18–316%) of the age-related target, respectively.

Laboratory Findings

Concentrations of plasma calcium (median, 2.40 mmol/L; range, 2.14–2.62 mmol/L), phosphate (median, 1.41 mmol/L; range, 0.77–1.69 mmol/L), alkaline phosphatase (median, 211 U/L; range, 54–651 U/L), parathormone (median, 45ng/L; range, 11–107 ng/L), and 25(OH)vitamin D (median, 48 nmol/L; range, 19–102 nmol/L) were all within the institutional reference values. The median concentration of serum IGF-1 and IGFBP-3 were 20 nmol/L (range, 9–69 nmol/L) and 3.5 mg/L (range, 1.8–5.3 mg/L), respectively; 1 patient had a serum IGF-1 level that was below the age-specific reference values.

Bone Mineral Density

Bone age corresponded (difference < ±1.0 year) to the chronologic age in 31 patients (70%). Bone age was advanced by >1.0 year in 3 patients and delayed by >1.0 year in 10 patients. The median Z-score at the lumbar spine, the total hip, and the whole body were −0.4 (range, −3.2 to +1.8), −0.4 (range, −5.0 to +1.7), and 0.0 (range, −2.3 to +2.6), respectively. A reduced BMD (Z score <−1.0) was observed at the lumbar spine in 9 patients (3 of whom were prepubertal), at the total hip in 13 patients (4 of whom were prepubertal), and in the whole body in 7 patients (1 of whom was prepubertal). In total, 16 of the 44 patients (36%) (10 boys [37%] and 6 girls [35%]) had a Z-score of <−1.0 at either the lumbar spine, the total hip, or the whole body. The Z-scores tended to be lower in pubertal and postpubertal patients compared with prepubertal patients (Fig. 1). This was evident especially at the hip, in which the median aBMD Z-score in prepubertal patients (−0.2; range, −0.5 to +1.7) was significantly higher than in pubertal patients (median, −1.1; range, −1.5 to +0.4) and postpubertal patients (median, −1.1; range, −2.6 to +0.5) (P = .03). The median aBMD at the total hip (P = .04) or whole body (P = .05) was already lower in pubertal patients, whereas decreased aBMD at the lumbar spine was evident in postpubertal patients (P = .07) compared with the prepubertal patients (Fig. 1).

Figure 1. Bone mineral density (BMD) in children and adolescents after allogeneic stem cell transplantation. BMD, expressed in Z-scores, is given separately in prepubertal (dashed bars), pubertal (white bars), and postpubertal (grey bars) subjects. The median, 25th, and 75th percentiles are indicated.

The patients with reduced or normal BMD did not differ with regard to the dietary yield of protein or vitamin D, physical activity, the cumulative dose of glucocorticoids or methotrexate, or hormonal insufficiencies (Table 2). The median serum levels of parathormone, calcium, and phosphate were found to be similar between the groups. Older age at the time of SCT was associated with reduced BMD (P = .03) (Table 2). There was a strong correlation found between chronic GVHD and reduced BMD (Table 2). Accordingly, 11 of the 15 patients with signs of chronic GVHD had a Z-score <−1.0. The median yield of calcium, expressed as a percentage of the age-related target level, was lower in those patients with a reduced BMD (median, 132%; range, 68–189%) compared with patients with a normal BMD (median, 220%; range, 58–319%) (P = .05).

Table 2. Clinical and Biochemical Parameters in SCT Patients With Normal and Reduced BMD (Presented as The Medians [Range])

	Normal BMD (n = 28)	Reduced BMD* (n = 16)	P
SCT indicates stem cell transplantation; BMD, bone mineral density (in Z-scores); ALL, acute lymphoblastic leukemia; SD, standard deviation; BMI, body mass index; GVHD, graft-versus-host disease; GH, growth hormone; PTH, parathormone. The Mann-Whitney U test was used in the comparisons. * Reduced BMD is defined as a Z-score of <−1.0 at the lumbar spine, total hip, or whole body. † The percentage of daily energy yield (recommendation of 10–15%). ‡ The percentage of recommended daily allowance (26 of 44 patients returned the dietary recall diary). § Training at least once a week (24 of 44 patients returned the questionnaire).
Clinical parameters
Diagnosis (n) (ALL vs others)	10/18	8/8	.36
Age at study, y	8.1 (4.7–22.6)	12.6 (5.4–20.4)	.006
Age at SCT, y	5.4 (1.0–15.9)	9.2 (1.4–18.6)	.03
Height, SD	−0.4 (−4.1–2.5)	−0.7 (−3.7–1.8)	.44
BMI, kg/m^2	16.8 (13.4–27.6)	16.4 (13.2–26.5)	.90
Prednisone
Cumulative dose, mg/kg	192 (0–1072)	201 (0–2175)	.68
Cumulative duration, mo	3 (0–24)	7 (0–24)	.22
Cumulative dose of methotrexate, mg	0 (0–45100)	3611 (0–45100)	.67
Duration of cyclosporine, mo	7 (0–21)	6 (0–27)	.58
Nutrition %
Protein†	16 (10–21)	18 (13–22)	.76
Vitamin D‡	103 (43–232)	108 (18–316)	.93
Calcium‡	220 (58–319)	132 (68–189)	.05
Physical activity§	13/17	5/7	.85
Chronic GVHD (n)	4	11	<.001
GH insufficiency (n)	2	1	.91
Hypogonadism (n)	3	4	.70
Biochemical parameters
Plasma calcium, mmol/L	2.39 (2.28–2.62)	2.45 (2.14–2.62)	.15
Plasma phosphate, mmol/L	1.42 (1.01–1.60)	1.27 (0.77–1.69)	.08
Plasma PTH, ng/L	48 (21–107)	42 (11–77)	.08
Plasma 25(OH) vitamin D, nmol/L	47 (19–102)	51 (26–84)	.29

Vertebral Fractures

IVA images for the assessment of vertebral morphology were available for all 44 patients. A total of 9 patients (20%) had abnormal vertebral morphology that was consistent with compression fractures in the thoracic and/or lumbar spine (Fig. 2). In 7 patients, these findings were verified by spinal radiographs; no radiographs could be obtained for the remaining 2 patients. Three additional patients had wedge-shaped vertebrae with an anterior compressions of 17%, 18%, and 19%; these were classified as normal but may represent an early stage of osteoporotic compression deformity.

Figure 2. Vertebral compression fractures in 3 patients. (A) A 6-year-old boy with multiple compression fractures affecting all thoracic and lumbar vertebrae 2 years after stem cell transplantation (SCT). (B) A 16-year-old boy with compression fractures at Th11, Th12, L1, and L2 3 years after SCT. (C) A 16-year-old girl with a compression fracture at Th 7 1.5 years after SCT.

In the 9 patients with vertebral fractures, a total of 31 compressed vertebrae were found. In 5 of these 9 patients, the fractures were contiguous (2–17 fractured vertebrae in each patient). Twenty of the 31 fractures (65%) were located in the thoracic spine and 11 fractures (35%) were located in the lumbar spine. In 7 of these 9 patients, the vertebral compression fractures were asymptomatic.

The patients with compression fractures did not differ from the others with regard to any transplantation-related, therapy-related, or dietary factors, nor with regard to demographic parameters (Tables 3 and Table 4). It is interesting to note that vertebral compressions did not appear to have any correlation with any of the BMD parameters (Table 3). Surprisingly, ALL was the underlying diagnosis in only 2 of the 9 patients with vertebral compressions compared with 16 of the 35 patients with normal vertebral morphology (P = .23).

Table 3. Clinical and Biochemical Parameters of SCT Patients With and Without Vertebral Fractures (Shown as The Medians [Range])

	Normal vertebrae (n = 35)*	Vertebral fractures (n = 9)†	P
SCT indicates stem cell transplantation; ALL, acute lymphoblastic leukemia; SD, standard deviation; BMI, body mass index; GVHD, graft-versus-host disease; GH, growth hormone; PTH, parathormone. The Mann-Whitney U test was used in the statistical analyses. * Height reductions of <20% were regarded as normal. † A height reduction of 20% in the anterior, middle, and/or posterior vertebral height was considered to be a vertebral compression fracture. ‡ Percentage of daily energy yield (recommendation of 10–15%). § Percentage of recommended daily allowance (26 of 44 patients returned the dietary recall diary). ‖ Training at least once a week (24 of 44 patients returned the questionnaire).
Clinical parameters
Diagnosis (n) (ALL vs others)	16/19	2/7	.29
Age, y	10.4 (4.7–22.6)	7.2 (5.3–17.6)	.30
Age at SCT, y	6.7 (1.4–18.6)	3.7 (1.0–14.8)	.18
Height, SD	−0.4 (−2.7–1.8)	−0.8 (−4.1–2.5)	.16
BMI, kg/m^2	16.8 (13.2–27.6)	17.4 (13.4–22.5)	.89
Prednisone
Cumulative dose, mg/kg	172 (0–1072)	242 (0–2175)	.41
Cumulative duration, mo	3 (0–24)	7 (0–23)
Cyclosporine, duration, mo	7 (0–27)	6 (1–18)	.69
Nutrition %
Protein‡	17 (10–22)	16 (13–18)	.45
Vitamin D§	100 (18–316)	130 (54–265)	.66
Calcium§	146 (58–319)	224 (122–252)	.31
Physical activity‖	14/19	4/5	.84
Chronic GVHD (n)	10	5	.22
GH insufficiency (n)	1	2	.39
Hypogonadism (n)	5	2	.73
Biochemical parameters
Plasma calcium, mmol/L	2.40 (2.24–2.62)	2.39 (2.14–2.55)	.59
Plasma phosphate, mmol/L	1.43 (0.77–1.69)	1.35 (1.15–1.58)	.16
Plasma PTH, ng/L	45 (13–99)	47 (11–107)	.47
Plasma 25(OH) vitamin D, nmol/L	48 (19–102)	49 (41–79)	.41
Bone mineral density
Lumbar spine, SD	−0.4 (−3.2–1.8)	−0.4 (−1.4–1.2)	.87
Total hip, SD	−0.4 (−2.6–1.7)	−0.9 (−5.0–1.3)	.37
Whole body, SD	0.0 (−2.3–2.6)	0.1 (−1.9–1.3	.87

Table 4. Clinical Characteristics of the Patients With Vertebral Fractures

Gender	Age, y	Diagnosis	Affected vertebrae	BMD,* Hip	BMD,* whole body	Cumulative dose of prednisone, mg/kg	Cumulative duration of prednisone, mo
BMD indicates bone mineral density; M, male; HLH, hemophagocytic lymphohistiocytosis; L, lumbar; ALL, acute lymphoblastic leukemia; T, thoracic; F, female; AML, acute myeloid leukemia; CML, chronic myeloid leukemia; LCH, Langerhans cell histiocytosis; WA, Wiscott-Aldrich syndrome. * Bone mineral densities are expressed in Z-scores.
M	6.2	HLH	L3-4	1.3	0.9	371	7
M	17.6	ALL	T11-12; L1-2	−2.0	−1.9	242	9
F	14.3	AML	T7	−1.0	−0.2	604	5
M	5.6	AML	T4	−0.9	0.4	39	2
M	5.3	AML	T8	−1.2	−0.8	0	0
F	16.5	CML	T7	−0.9	−0.1	26	1.5
M	7.2	LCH	T5; L2-3	0.6	1.3	483	23
M	6.6	WA	all T + L vertebrae	−5.0	–	2175	18
F	9.0	ALL	T7	−1.8	−1.5	119	20

DISCUSSION

The results of the current study demonstrated a reduced BMD in 36% of the 44 children who underwent SCT who were examined for various parameters of bone health ≥6 months after undergoing SCT. Furthermore, a significant proportion of the study subjects had symptomatic osteoporosis with fractures. To our knowledge, this is a novel observation in SCT patients. These transplanted children had several risk factors that possibly contributed to the development of osteopenia or osteoporosis. These include glucocorticoid therapy of the underlying disease and GVHD, immobilization, and several endocrine disturbances, including delayed growth7, 11 and hypogonadism.7 Previous studies suggest that approximately one-third of patients who undergo SCT in childhood develop at least osteopenia.7, 18 However, the data regarding BMD are controversial.7, 11 Nysom et al reported only slightly reduced whole body bone mass in adolescents at a median of 8 years after allogeneic SCT.11 In contrast, we observed a substantially reduced BMD (Z-score of <−1.0 at the lumbar spine or hip) in 38% of long-term SCT survivors in young adulthood.7 Similarly, decreased BMD for age has been reported by Bhatia et al18 and Daniels et al.19 Compromised bone health is not only a concern in the adult SCT survivors but also in recently transplanted children, as evidenced by the significant disturbances in various skeletal parameters noted in the relatively young (median age, 10 years) current study population.

The most concerning observation in the current study is that 20% (9 of 44 patients) of all the study patients already had vertebral compression fractures at a young age. The vertebral fractures were not associated with overweight or any other demographic parameter or any of the biochemical parameters studied. The heights in those patients with and those without vertebral compression fractures did not appear to differ significantly, most likely because several factors in addition to vertebral crush fractures may impact growth in the posttransplant period. Seven of the 9 patients with compression fractures were asymptomatic. Asymptomatic vertebral fractures also have been reported in a pediatric solid organ transplantation population19 and adult cardiac transplantation recipients.20 However, it is possible that the true prevalence of compression fractures may be even higher because IVA images may not be as accurate as radiographs in diagnosing vertebral fractures in children. Without a broad and active survey of bone health, asymptomatic vertebral fractures easily may go unrecognized in SCT patients. The fractures were located for the most part in the thoracic region, which is also the most typical location for traumatic vertebral fractures21 and osteoporotic compression fractures in adults.22 In addition, the majority of the affected children had multiple fractures affecting adjacent vertebral bodies.

The high frequency of vertebral fractures in this population should have at least 3 clinical implications. First, it is important to include assessment of the spine in the follow-up of even young survivors of SCT. Second, finding 1 compressed vertebral body should result in a careful examination of the entire spine. Third, it should be kept in mind that normal BMD does not rule out compression fractures. Instead, no correlation was noted between reduced BMD and vertebral fractures in the current study. This may be due to a long interval between the fracture and BMD measurement or due to differences in the location of the fractured vertebrae (usually the thoracic spine) and BMD assessment (lumbar spine or proximal femur). Furthermore, crushed vertebrae are more dense than normal vertebral bodies and compression fractures may result in falsely normal BMD readings, especially when located in the BMD measurement area at the lumbar spine. Previous studies have shown that DXA is of limited value in the assessment of bone quality.23 However, spinal imaging does not reduce the need to evaluate BMD in SCT patients. We do not yet know what, in the long term, will be the overall risk of fractures in these patients. Therefore, recognizing the individuals with reduced BMD will most likely result in the enhancement of predictive measures. Taken together, we are convinced that spinal imaging adds important clinically relevant information to the evaluation of the bone health in children who have undergone SCT.

Chronic GVHD is likely to play a significant role in the development of posttransplant osteoporosis. In the current study, GVHD was found to be strongly correlated with reduced BMD (ie, a Z-score <−1.0 at the lumbar spine, total hip, and/or whole body). Stern et al reported decreasing BMD values during the treatment of chronic GVHD in a small series of adult SCT patients.24 Cyclosporin and glucocorticoids, used in the prophylaxis and therapy of GVHD, are known risk factors for osteoporosis.25 To our knowledge, the role of GVHD per se in bone metabolism has not yet been established. However, some of the cytokines that are active in the GVHD process26 are also effectors in bone metabolism. Interleukin (IL) −1, IL-6, and tumor necrosis factor-α are known to stimulate osteoclast activity and increase bone resorption.27 In adult patients, immunosuppression, the presence of GVHD, and increased cytokines in the early post-SCT period were found to be associated with later bone loss.28

To our knowledge, the current study population was younger than those in other studies addressing bone health after SCT in childhood.7, 18, 19, 29 This enabled us to examine bone health across different stages of puberty. In prepubertal children, the median BMD was within normal range at all measured sites. In those patients who advanced to Tanner stages 2–3, the BMDs at the total hip and whole body tended to be lower (Fig. 1). The postpubertal children (Tanner stages 4–5) were found to have significantly lower lumbar spine BMDs than children at the earlier stages of puberty. Normal pubertal development with increased sex steroid production and growth hormone secretion is essential for bone mass accrual; in optimal conditions, skeletal mass doubles during the pubertal years.30 The results of the current study suggest that the underlying chronic illness and the SCT performed during the period of skeletal growth constitute a significant threat to normal bone mass development and the acquisition of an optimal peak bone mass. We speculate that hormonal factors might participate as effectors in the process. Both growth hormone31 and sex steroids30 are needed in skeletal maturation (ie, for growth in height and increase in bone mass). They are known to act on both osteoblasts and osteoclasts with a net effect of bone accumulation.30, 32 Clinically, growth hormone insufficiency is associated with reduced BMD in both children30, 33 and adults.33 In addition, a delay of menarche until age ≥16 years increases the risk of vertebral fractures in later life.34 Previous studies have suggested that delayed pubertal growth and reduced height for age,7, 19 as well as puberty-onset hypogonadism,7 are all associated with impaired bone health in young adulthood after SCT is performed in childhood. Due to the small number of patients with hormonal insufficiencies, the etiology could not be addressed in the current study. Consequently, there is a need for a prospective follow-up study of SCT recipients to establish the role of hormonal factors in pubertal bone maturation.

The yield of calcium and vitamin D were comparable in patients with normal and those with reduced BMD. Furthermore, there was no difference in physical activity observed between the groups. The low number of returned questionnaires may have influenced this result. However, due to a substantially increased risk of osteoporosis in young SCT patients, it is worthwhile to be alerted to avoid further risks (ie, nutritional deficiencies, immobilization, and smoking) in later life.

The current study was comprised of a relatively large cohort of children who had undergone SCT and whose bone health was thoroughly evaluated for various parameters, including vertebral morphology. The study cohort was representative compared with those not willing to participate. However, the cross-sectional design limited the interpretation of some findings. Without longitudinal data, no conclusions could be made regarding the timing of the fractures and it is possible that some of these fractures developed during the pretransplantation period, especially in patients with leukemia and lymphoma. The evaluation of BMD changes during puberty and the possible hormonal insufficiencies involved in particular would require a longitudinal approach. In interpreting the DXA images, special attention was paid to exclude potential sources of error. The choice of wrong reference data may lead to misinterpretation of the BMD measures.35 The BMD reference data for white children in the U.S., provided by the manufacturer, was considered appropriate because first, the heights of white children in the U.S. do not differ significantly from the heights of Finnish children of the same age and gender15, 36 and second, because BMD and its variations in young Finnish adults is similar to that in the U.S. white reference population (unpublished data). All BMD Z-scores were corrected for skeletal maturity; therefore, possible differences in pubertal timing between the reference population and the study population would not cause errors in data interpretation. In addition, for those patients with compression fractures in the lumbar spine, only total hip and whole-body BMD assessments were included in the analyses. However, even after controlling for these factors, the DXA data remained of limited value and provided no information regarding bone quality. More accurate methods to assess bone quality, such as quantitative computed tomography and bone biopsy, and longitudinal evaluations of a larger cohort of patients are needed in future studies.

Various medications, including calcium substitution alone or in combination with calcitonin4 or with vitamin D and estrogen37 have proven to be ineffective in the prevention of bone loss in adult SCT recipients. More promising results were obtained by combining bisphosphonate (pamidronate) with calcium, vitamin D, and sex hormone replacement therapy; such treatment reduced bone resorption but did not appear to completely normalize bone turnover or prevent bone loss in SCT patients.38 Bisphosphonate therapy also has been used successfully in children.39 However, the therapeutic approach in children should always begin with the correction of nutritional and hormonal deficiencies, including sex steroids and growth hormone. Bisphosphonates should be used cautiously and only by physicians familiar with this treatment.

In conclusion, greater than one-third of children and adolescents were found to have decreased BMD after undergoing SCT in childhood. Bone mass accrual was found to be compromised, especially during the pubertal years, and GVHD was a strong risk factor. Asymptomatic spinal compression fractures were frequently present. Although children undergoing SCT are at a significant risk of developing osteoporosis and compression fractures; their bone health should be carefully monitored during the posttransplantation follow-up to allow for the early initiation of bisphosphonate therapy when indicated.

Acknowledgements

We thank Päivikki Rissanen, RN, for technical assistance.