Predictors of bony morbidity in children with acute lymphoblastic leukemia

BACKGROUND: To evaluate the relationship between lumbar spine (LS) bone mineral density (BMD) and patient-, disease-, and therapy-related variables, and to define the risk-factors for fractures in children receiving therapy on Dana-Farber Cancer Institute acute lymphoblastic leukemia (ALL) protocols. METHODS: Children (≤18 years) diagnosed with ALL during the period 1995-2006, who are in first clinical remission, were included (n = 124). Dual-energy X-ray absorptiometry provided LS-BMD at diagnosis (n = 46) and during continuation therapy. LS-BMD was expressed as Z scores based on local population norms. Regression analyses evaluated the risk of osteopenia (Z-score -1.01 to -1.99, osteoporosis (Z-score -2.00 or less) and fractures. RESULTS: At diagnosis, 14 0f 46 (30%) patients had osteopenia and 5 (11%) had osteoporosis; whereas, during continuation therapy, 47 of 124 (39.5%) patients had osteopenia, and 10 (8%) had osteoporosis. LS-BMD at diagnosis had a positive linear relationship with LS-BMD during continuation therapy (Pearson correlation coefficient 0.619, P < 0.0001). Multivariable analyses identified age ≥10 years and LS-BMD at diagnosis as independent predictors of LS-BMD during continuation therapy. Twenty-three (18.5%) patients developed fractures. Dexamethasone therapy (OR 3.4, 95% CI 1.31, 7.52, P = 0.01) and lower LS-BMD during the continuation therapy (OR 1.8, 95% CI 1.2, 2.8, P = 0.01) were independent predictors of fracture. CONCLUSIONS: Older age and lower LS-BMD at diagnosis are predictors of lower LS-BMD during continuation therapy. Dexamethasone and lower LS-BMD during continuation therapy are associated with fractures. Using these variables it is feasible to develop a predictor model to define the risk of bony morbidity in children receiving ALL therapy.