Identification of vertebral fracture and non-osteoporotic short vertebral height in men: The MrOS study

Lynne Ferrar, Guirong Jiang, Peggy M. Cawthon, Ria San Valentin, Robin Fullman, Lori Lambert, Steven R. Cummings, Dennis M. Black, Eric Orwoll, Elizabeth Barrett-Connor, Kris Ensrud, Howard A. Fink, Richard Eastell

Research output: Contribution to journalArticlepeer-review

65 Scopus citations


Non-osteoporotic SVH may mimic VF but is excluded in ABQ. In men, this led to discordance between ABQ and other methods, but SVH was not linked to low bone density. Exclusion of SVH could reduce false positives. Introduction: Non-osteoporotic short vertebral height (SVH) may mimic vertebral fracture (VF). The aims were to (1) compare the prevalence of VF in elderly men using the algorithm-based qualitative (ABQ), semiquantitative (SQ), and triage-quantitative morphometric (triage-QM) methods; (2) identify reasons for discordance between methods; and (3) determine whether SVH identified by ABQ is linked to low BMD. Materials and Methods: We studied a subset of 732 men ages ≥65 yr participating in the Osteoporotic Fractures in Men (MrOS) Study. Criteria for VF were (1) ABQ: endplate depression; (2) SQ: estimated vertebral height reduction ≥20%; (3) triage-QM: vertebral height ratio >3 SD below the reference mean, on radiographs showing evidence of VF. Criteria for SVH (ABQ) were apparent "reduction" in vertebral height ≥ -15%, without evidence of endplate depression. Results: The prevalence of at least one VF was 10% (ABQ); 13% (SQ) and 11% (QM-triage) and of at least one SVH (ABQ) was >50%. Agreement between methods was moderate (κ = 0.42-0.62). Discordance between methods related mainly to classification of mild thoracic wedging or possible traumatic VF by ABQ. Mean BMD was lower in men with VF (any diagnostic method) than in those without (two-sample t-test, p < 0.05). For ABQ, BMD was similar in men with SVH (no VF) and men with normal vertebrae (ANOVA, p > 0.05). Mean BMD was significantly lower than expected in 40 men with VF identified by all three methods and average or more than average in those identified by a single method. Conclusions: Among elderly men (1) the prevalence of VF ranges from 10% to 13%: (2) agreement between diagnostic methods is moderate: discordance relates mainly to differential classification of mild thoracic deformities or ABQ definition of VF as traumatic; and (3) SVH identified by ABQ is common and not linked to low BMD.

Original languageEnglish (US)
Pages (from-to)1434-1441
Number of pages8
JournalJournal of Bone and Mineral Research
Issue number9
StatePublished - Sep 2007

ASJC Scopus subject areas

  • Endocrinology, Diabetes and Metabolism
  • Orthopedics and Sports Medicine


Dive into the research topics of 'Identification of vertebral fracture and non-osteoporotic short vertebral height in men: The MrOS study'. Together they form a unique fingerprint.

Cite this