Oral Presentation ANZBMS-MEPSA-ANZORS 2022

Agreement and associations between bone variables assessed via standard and high-resolution peripheral quantitative computed tomography (#91)

Jakub Mesinovic 1 2 , David Scott 1 2 , Cat Shore-Lorenti 2 , Roger Zebaze 2 , Mícheál Ó Breasail 3 , Camelia Lim 2 , Zihui Ling 2 , Peter Ebeling 2 , Ayse Zengin 2
  1. Institute for Physical Activity and Nutrition (IPAN), Deakin University, Burwood
  2. Department of Medicine, Monash University, Clayton
  3. Population health Sciences, University of Bristol, Bristol

Introduction: Comparisons of bone variables measured by two commonly used three-dimensional imaging modalities, peripheral quantitative computed tomography (pQCT) and second-generation high-resolution pQCT (HR-pQCTII), are limited. We assessed agreement and associations between volumetric bone mineral density (vBMD) and bone strength parameters at the radius and tibia via pQCT and HR-pQCTII.

Methods: Thirty older adults aged >50 years were recruited. Total, trabecular and cortical vBMD assessed at the radius and tibia by pQCT (both: 4%; radius: 33%; tibia: 38%) and HR-pQCTII (both: 4% and 30%). Stress-strain index (SSI) and bone stiffness were assessed at the proximal radius and tibia via pQCT and HR-pQCTII, respectively. Bland-Altman plots assessed agreement between pQCT and HR-pQCTII parameters. Pearson correlations and linear regression explored relationships between bone variables and proportional bias, respectively.

Results:

6296fd27243f5-ANZBMS+Abstract+Table.jpg

There was systematic bias (range: 9-26%) between pQCT and HR-pQCTII vBMD variables at the radius and tibia (all P<0.001; Table). Proportional bias was not observed between absolute vBMD measures at any site (all P>0.05) but was observed for percentage mean differences in total and trabecular vBMD at the distal radius and tibia (all P<0.05); participants with lower vBMD had higher percentage mean differences. pQCT- and HR-pQCTII-determined total, trabecular and cortical vBMD at the radius were strongly correlated (all r>0.8 and P<0.001). SSI was strongly correlated with bone stiffness at the proximal radius (r=0.93) and tibia (r=0.93).

Conclusion: Systematic bias exists between pQCT- and second-generation HR-pQCT-determined vBMD measurements at the radius and tibia, so these imaging modalities cannot be used interchangeably. Proportional bias in total and trabecular vBMD percentage mean differences suggests bias is more pronounced in those with poorer bone health. Although there was poor agreement between pQCT and HR-pQCTII, bone density and strength estimates derived via both imaging modalities were strongly correlated and likely have similar fracture prediction capabilities.