Abstract
Background : Reduced bone mineral density (BMD) has been reported in adults and children with sickle cell anemia (SCA). Dual energy x-ray absorptiometry (DXA) is routinely used for measuring BMD because of less radiation exposure and lower cost. However, changes in vertebral body shape, marrow hyperplasia and bone infarction due to SCA may affect the evaluation of BMD with DXA. Hence, we compared DXA with quantitative computerized tomography (QCT), which measures true volumetric density, and may be less influenced by bone changes.
Methods : The study enrolled children between 9–19 years of age with SCA, and one or more severe manifestations: >2 hospital admissions/year, growth failure, avascular necrosis, or regular red cell transfusions for sickle cell-related complications. BMD of lumbar spine was determined by performing DXA of lumbar spine (Hologic Delphi-A, Bedford, MA). The apparent volumetric bone mineral density (BMAD) was calculated from bone mineral content, and compared to age, sex and ethnicity-matched reference data. BMD of the lumbar spine was also measured by QCT (Mindways Software, San Francisco, CA), and compared to age and sex-appropriate reference data.
Results : The study has enrolled 25 patients (13 females and 12 males), of which 16 were younger than 14 years. In 6 children the height was <10th centile for age. Thirteen patients were on regular transfusions for >6 months, including 10 who had been transfused for >2 years. Calcium intake, assessed by a standardized questionnaire, was less than recommended dietary allowance in 13 patients.
The z-score for BMAD determined by DXA was > −1.0 in 8, between −1.0 and −2.0 in 5, and < −2.0 in 12 patients. The z-score for lumbar spine by QCT was > −1.0 in 20, between −1.0 and −2.0 in 1 and < −2.0 in 4 patients. DXA-derived BMD (areal density) and BMAD (apparent volumetric density) z-scores did not differ significantly (p=0.16). On the other hand, the paired values of z-scores by DXA (BMAD) and QCT were significantly different (p=.002). When z-scores were categorized as greater or less than −1.0, the results were concordant in 13 (both DXA and QCT normal in 8, and both DXA and QCT abnormal in 5), and discordant in 12 cases (abnormal DXA with normal QCT in every case). Among patients in discordant group, 9/12 had been on regular red cell transfusions for >6 months, compared to 4/13 with concordant results (p=.047). There was no difference in the serum ferritin values between the two groups (p=.685).
No significant difference in the prevalence of low BMAD z-scores was detected between groups based upon age, calcium intake, or growth failure. Five out of the 12 patients with BMAD z-score < −2.0 were not on regular transfusion program.
Conclusions : Almost half of the children with SCA had BMD below −2 standard deviations compared to age-matched controls. Low BMD was observed in chronically transfused as well as non-transfused children. In comparison, 16% of the patients were classified as low BMD (z < −2.0) by QCT. The paired DXA/QCT results were discordant in half of the sample, with patients on regular transfusions for >6 months more likely to have normal QCT results. It is likely that the reduction in marrow hyperplasia following initiation of regular transfusions may disproportionately affect the trabecular BMD measured by QCT. Longitudinal evaluation of BMD in patients starting on transfusion program could help to define the effect of transfusions on measures of BMD in SCA.
Author notes
Corresponding author
