A Validated Risk Prediction Model for Bone Fragility in Children with Acute Lymphoblastic Leukemia

Introduction

Due to bone fragility, children with acute lymphoblastic leukemia (ALL) have a 6-fold greater fracture risk during therapy compared to peers. Osteoporotic fractures are a concern, as they lead to adverse health outcomes including pain, loss of height due to vertebral deformity, and (transient) disability. In previous studies, lower lumbar spine bone mineral density (LS BMD) at ALL diagnosis was found to be prognostic for the occurrence of future fractures. However, routinely performing dual-energy X-ray absorptiometry (DXA) in each newly diagnosed child is not universally feasible. The aim of this study is to develop and validate an easy to use clinical risk prediction model for low lumbar spine bone mineral density (LS BMD Z-score ≤-2.0) at diagnosis, as an important indicator for fracture risk and further treatment-related BMD aggravation.

Methods

Children treated for ALL according to the Dutch Childhood Oncology Group (DCOG-ALL9; model development) protocol (n=249; median age: 7.6 years [range: 4.0-16.6 years]) and children from the Canadian STeroid‐Associated Osteoporosis in the Pediatric Population (STOPP; model validation) cohort (n=99; median age: 7.3 years [range: 4.0-16.6 years]) were included in this study.

Multivariable logistic regression analyses were used to develop the prediction model for low LS BMD at diagnosis, defined as a Z-score ≤-2.0 (evaluated with DXA). Candidate predictors included sex, age, height and weight Z-scores at diagnosis of ALL. The receiver operating characteristic area under the curve (AUC) was assessed for model performance.

To confirm the association between low LS BMD at diagnosis and bone fragility during and shortly following ALL therapy, we performed multivariable logistic regression analyses. The dependent variables were: one or more symptomatic fractures from ALL diagnosis to 12 months following treatment cessation and low LS BMD at cessation of treatment.

In addition, because of homogeneity in the intended glucocorticoid doses, we combined data from the DCOG-ALL9 and STOPP cohorts and performed multivariable pooled cohort analyses (meta-analysis). Potential associations between the six-month cumulative glucocorticoid dose and fractures that occurred in the first year of therapy, were explored. Furthermore, we assessed potential associations between the cumulative glucocorticoid dose at cessation of therapy, and the endpoints 'low LS BMD at therapy cessation' and 'fractures that occurred during treatment and within 12 months following treatment cessation'.

Results

The prediction model for low LS BMD at diagnosis included weight Z-scores (β = -0.70) and age (β = -0.10) at diagnosis. This model had an AUC of 0.71 (0.63 to 0.78) in the DCOG-ALL9 cohort, and resulted in correct identification of 71% of patients with low LS BMD at ALL diagnosis. Validation on the STOPP cohort showed an AUC of 0.74 (95% CI = 0.63 to 0.84). To calculate the probability of low LS BMD at ALL diagnosis for an individual patient, an online calculator is available at http://lsbmd-risk-calculator.azurewebsites.net/

We confirmed that low LS BMD at diagnosis is associated with LS BMD at treatment cessation (OR = 5.9; 95% CI = 3.2 to 10.9) and with symptomatic fractures (OR = 1.7; 95% CI = 1.3 to 2.4) that occurred from diagnosis until 12 months following treatment cessation.

In pooled meta-analysis, lower LS BMD at diagnosis (OR = 1.6, 95% CI = 1.1 to 2.4) and six-month cumulative glucocorticoid dose (OR = 1.9, 95% CI = 1.1 to 3.3, for every gram increase) were associated with symptomatic fractures that occurred in the first year of therapy. Higher cumulative glucocorticoid dose at cessation of therapy (OR = 1.5, 95% CI = 1.2 to 2.0, for every gram increase), lower LS BMD Z-scores at diagnosis (OR = 7.9, 95% CI = 4.8 to 13.1) and higher age at diagnosis (OR = 1.6, 95% CI = 1.4 to 1.8), were associated with low LS BMD at cessation of therapy.

Conclusion

We developed and successfully validated a risk prediction model for low LSBMD at diagnosis in children aged 4-18 years with ALL. This is important because low LS BMD at diagnosis was strongly associated with bone fragility and fractures during and shortly following treatment for ALL. Our easy to use prediction model, can facilitate awareness and early identification of bone fragility in individual pediatric ALL patients, without performing DXA examination.