Modifiable Risk Factors Are Associated with Reduced Bone Mineral Density and Fractures in a National Cohort of 2,003 Dutch Childhood Cancer Survivors

Introduction Childhood cancer survivors are at risk of developing skeletal late effects. However, evidence on risk factors for very low bone mineral density (BMD, Z-score ≤-2), as well as on risk and risk factors of (vertebral) fractures is limited. We investigated this in a national cohort of Dutch childhood cancer survivors treated from 1963-2002.

Methods This cross-sectional study is part of the Dutch Childhood Cancer Survivor Study (DCCSS) LATER cohort. Individuals who survived at least five years after cancer diagnosis and aged 18-45 years at invitation were included. BMD was assessed by dual-energy X-ray absorptiometry (DXA, n=1,548). We defined low BMD as a Z-score ≤-1 and very low BMD as a Z-score ≤-2. Fractures that occurred >5 years after cancer diagnosis were assessed by medical history (n=1,892). Fracture incidence was compared with Swedish normative data by calculating a standardized incidence ratio (SIR). Vertebral fractures were evaluated by DXA-assisted vertebral fracture assessment using the Genant semiquantitative method (n=249). Associations between demographic, treatment-related, endocrine, as well as lifestyle-related factors and reduced BMD and (vertebral) fractures were evaluated using univariable or multivariable logistic regression models.

Results A total of 2,003 childhood cancer survivors participated in this study (32.3% lymphoid leukemia, 5.1% myeloid and other leukemia, 18.6% lymphoma, 28.0% solid tumors, and 16.0% CNS and brain tumors) with a mean age of 33.1 (standard deviation ±7.2) years. Low BMD at any site occurred in 35.3% (95%CI=32.1-38.5%) of survivors with hematologic malignancies, and very low BMD in 9.6% (95%CI=7.7-11.2%). Additionally, 32.3% (95%CI=29.5-35.2%), 16.8% (95%CI=14.6-19.2%), and 13.3% (95%CI=11.3-15.5%) of survivors with hematologic malignancies had experienced any fracture, a long bone fracture, or a fragility fracture, respectively. The SIR of any first fracture that occurred at least 5 years after cancer diagnosis was 3.53 (95%CI=3.06-4.06) for male survivors and 5.35 (95%CI=4.46-6.52) for female survivors. Vertebral fractures were prevalent in 13.3% of evaluable survivors. In the models for low or very low BMD, male sex, underweight, shorter follow-up time (continuous), total body irradiation, cranial irradiation, carboplatin (≥2,000 mg/m²), alkylating agents (≥8,000 g/m²), hypogonadism, growth hormone deficiency (GHD), hyperthyroidism, low physical activity, severe vitamin D deficiency, vitamin B12 deficiency, and folic acid deficiency were statistically significant. Male sex, obesity, previous/current smoking, and very low lumbar spine BMD were significantly associated with reported clinical fractures. Older attained age, platinum compounds, GHD, and low physical activity were significantly associated with vertebral fractures.

Conclusions: Childhood cancer survivors are at increased risk of any fracture. Reduced BMD at follow-up (especially very low lumbar spine BMD) was significantly associated with fractures, which underscores the importance of active BMD surveillance for high-risk survivors (i.e. those treated with cranial/craniospinal or total body irradiation). We identified high-dose carboplatin as a new treatment-related risk factor for low BMD, whereas corticosteroid (dose) was not significantly associated with reduced BMD or fractures. In addition, several modifiable risk factors for reduced BMD (i.e. hypogonadism, GHD, low physical activity, severe vitamin D deficiency, vitamin B12 deficiency, and folic acid deficiency) and vertebral fractures (i.e. GHD and low physical activity) were identified.

No relevant conflicts of interest to declare.