Rate of bone density change does not enhance fracture prediction in routine clinical practice

J Clin Endocrinol Metab. 2012 Apr;97(4):1211-8. doi: 10.1210/jc.2011-2871. Epub 2012 Jan 25.


Context: There is contradictory information on whether the rate of bone mineral density (BMD) loss is an independent risk factor for osteoporotic fractures and whether this should be included in fracture prediction systems.

Objective: This study was undertaken to better define rate of BMD loss as a contributor to fracture risk in routine clinical practice.

Design and setting: We performed a retrospective cohort study using a database of all clinical BMD results for the province of Manitoba, Canada.

Patients: We included 4498 untreated women age 40 yr and older at the time of a second BMD test performed between April 1996 and March 2009.

Main outcome measures: A total of 146 women with major osteoporotic fracture outcomes after the second BMD test (mean observation, 2.7 yr) and relevant covariates were identified in population-based computerized health databases.

Results: Annualized percentage change in total hip BMD was no greater in fracture compared to nonfracture women (-0.4 ± 1.7 vs. -0.5 ± 1.4; P = 0.166). After adjustment for final total hip BMD, other covariates, and medication use, rate of total hip BMD change did not predict major osteoporotic fractures (hazard ratio, 0.95 per sd decrease; 95% confidence interval, 0.81-1.10). Similar results were also seen in analyses based upon change in lumbar spine and femoral neck BMD.

Conclusions: We found no evidence that BMD loss, as detected during routine clinical monitoring, was a significant independent risk factor for major osteoporotic fractures.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Absorptiometry, Photon
  • Adult
  • Aged
  • Aged, 80 and over
  • Biomarkers
  • Bone Density*
  • Cohort Studies
  • Databases, Factual
  • Female
  • Femur Neck
  • Hip
  • Humans
  • Kinetics
  • Lumbar Vertebrae
  • Manitoba / epidemiology
  • Medical Record Linkage
  • Middle Aged
  • Osteoporotic Fractures / epidemiology*
  • Proportional Hazards Models
  • Retrospective Studies


  • Biomarkers