Predictors of bone mineral density in human immunodeficiency virus-1 infected children

J Pediatr Gastroenterol Nutr. 2005 Sep;41(3):339-46. doi: 10.1097/01.mpg.0000174468.75219.30.


Objectives: To compare bone mineral density (BMD) among human immunodeficiency virus (HIV)-infected children with population norms and to determine predictors of BMD in HIV-infected children.

Methods: Total body BMD was measured by dual energy x-ray absorptiometry in 37 HIV-infected children and nine sibling controls at baseline. Clinical, dietary and anthropometric data were obtained at the time of the dual energy x-ray absorptiometry examination. Age- and gender-adjusted z scores were calculated for BMD, body mass index, weight and height from population standards. Age-adjusted percentiles were determined for dietary intake of calcium and vitamin D. Differences in BMD z scores between HIV-infected children and sibling controls were determined and adjusted for height and weight, as were independent risk factors for lower BMD among infected children. Eighteen HIV-infected children and 5 controls had serial BMD measures.

Results: Compared with population norms, HIV-infected children had significantly lower BMD z scores (-0.51 SD, P = 0.004), in contrast with controls who had normal z scores (0.38 SD, P = 1.0). However, there was no difference in BMD z scores between HIV-infected children and the small number of sibling controls, adjusted for height and weight. Among HIV-infected children, lower BMD z scores were independently associated with lower weight z scores (P < 0.0001), lower height z scores (P = 0.01), advanced (stage B or C) HIV stage (P = 0.01) and age greater than 8 years (P < 0.0001). In the same model, multivitamin use (P = 0.009) and African American race (P = 0.001) were associated with better BMD z scores, with nevirapine use showing borderline positive effect (P = 0.06). All results were adjusted for Tanner stage. Change in BMD z score over time showed that there was no change or an increase in BMD in 100% of controls but in only 44% of the HIV-infected children (P = 0.09).

Conclusion: When compared with population norms, HIV-infected children had lower than expected bone mass for their age and gender that may be attributable to delays in growth, sexual maturity, time (length of HIV infection), ethnicity and disease severity. Dietary intake of calcium and vitamin D were not associated with bone loss, but most children had suboptimal intake. However, multivitamin use was strongly associated with better bone mineral density.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Absorptiometry, Photon / methods
  • African Americans / statistics & numerical data
  • Antiretroviral Therapy, Highly Active / adverse effects
  • Body Composition
  • Body Height / physiology
  • Body Mass Index
  • Body Weight / physiology
  • Bone Density / drug effects
  • Bone Density / physiology*
  • Case-Control Studies
  • Child
  • Child Development*
  • Cross-Sectional Studies
  • European Continental Ancestry Group / statistics & numerical data
  • Female
  • HIV Infections / drug therapy
  • HIV Infections / physiopathology*
  • HIV-1*
  • Hispanic Americans / statistics & numerical data
  • Humans
  • Longitudinal Studies
  • Male
  • Predictive Value of Tests
  • Prospective Studies
  • Severity of Illness Index
  • Vitamins / administration & dosage


  • Vitamins