Skip to main navigation Skip to search Skip to main content

Relationship between trabecular vertebral body density and fractures: A quantitative definition of spinal osteoporosis

  • Clarita V. Odvina
  • , Jon E. Wergedal
  • , Cesar R. Libanati
  • , Eloy E. Schulz
  • , David J. Baylink

Research output: Contribution to journalArticlepeer-review

Abstract

To evaluate the relationship between vertebral fractures and trabecular vertebral body density (TVBD), as measured by computed tomography (CT), we evaluated 110 female and 38 male patients referred consecutively to our clinic for an osteoporosis evaluation. Number of fractures per patient and TVBD was negatively correlated in both males and females (r = -.64, P < .001 and r = -.69, P < .001, respectively). Based on this relationship and that between percent of patients with fracture and TVBD, we devised four different approaches to calculate the fracture threshold. (1) Because the x-axis intercept of this regression line represents the TVBD value at zero fractures, this intercept can be considered the fracture threshold, which was 103 mg/cm3 for females and 132 mg/cm3 for males. (2) Breakpoint analysis of the relationship between the number of vertebral fractures per patient v TVBD gave a fracture threshold value of 98 mg/cm3 for females, but for males we were unable to compute a threshold value because of the small sample size. The percentage of patients with fractures was also negatively correlated with TVBD for males and females (r = -.98, P < .001, and r = -.94, P < .001, respectively). (3) the x-axis intercept of this relationship, which represents the fracture threshold, was 123 mg/cm3 for males and 101 mg/cm3 for females. (4) The fracture threshold, calculated as the mean TVBD + 2 SD for patients with fracture(s), was 120 and 92 mg/cm3 for males and females, respectively. In females, because the fracture threshold values estimated by these four different methods are in close agreement, and because the fracture threshold of 98 mg/cm3 determined by breakpoint analysis is intrinsically the most reliable, we suggest that an operational fracture threshold be assigned a value of 100 mg/cm3 and that osteoporosis be quantitatively defined as a TVBD value of 100 mg/cm3 or less. In males, although the fracture threshold was not as well defined, the values obtained by the different methods were in close agreement (range, 120 to 132 mg/cm3); also, the fracture threshold was higher for males than females (eg, 123 ± 7 v 101 ± 2 mg/cm3, P < .001). In females, TVBD was inversely related to age, but the fracture threshold was not affected by age (from 38 to 88 years) indicating that our fracture threshold is applicable at all ages and that age does not have an effect on fracture risk independent of its effect on TVBD.

Original languageEnglish
Pages (from-to)221-228
Number of pages8
JournalMetabolism
Volume37
Issue number3
DOIs
StatePublished - Mar 1988

ASJC Scopus Subject Areas

  • Endocrinology, Diabetes and Metabolism
  • Endocrinology

Cite this