Poster Presentation ESA-SRB-ANZBMS 2024 in conjunction with ENSA

Automated semi-quantitative abdominal aortic calcification scores are associated with incident atrial fibrillation and flutter: Results from the UK Biobank (#382)

Alexander J. Rodriguez 1 , Marc Sim 2 , James Webster 3 , Cassandra Smith 2 , Afsah Saleem 2 , Syed Zulqarnain Gilani 2 , John Kemp 4 , Nicholas C. Harvey 5 , John T. Schousboe 6 , Joshua R. Lewis 2
  1. Bone and Muscle Research Group, School of Clinical Sciences at Monash Health, Monash University, Melbourne, Victoria, Australia
  2. Institute for Nutrition Research, School of Medical and Health Sciences, Edith Cowan University, Joondalup, Western Australia, Australia
  3. Applied Health Research Unit, Nuffield Department of Population Health, Oxford University, Oxford, UK
  4. Mater Research Institute, University of Queensland, Brisbane, Queensland, Australia
  5. MRC Lifecourse Epidemiology Centre, University of Southampton, Southampton, UK
  6. Park Nicollet Clinic, HealthPartners Institute, Minneapolis, USA

Background: Abdominal aortic calcification (AAC) represents advanced atherosclerotic plaques and has been associated with poor prognosis. AAC stiffens the aorta which is hypothesised to increase systemic afterload, promote atrial remodelling, fibrosis, and heart strain; overall increasing the risk for atrial fibrillation (AF). No previous study has investigated if the presence of AAC is associated with incident AF or atrial flutter (AF/F)

Methods: Lateral spine images from GE iDXA in participants from the UK Biobank Imaging Study (2014-2022) were read by a validated machine learning tool to automatically estimate AAC based on a semi-quantitative 24-point scale (ML-AAC24)1. Incident AF/F was obtained through record linkage within the UK Biobank. Cox proportional hazards models were used to estimate the risk of incident AF/F in those with severe AAC (scores 6+) compared to those with moderate (2-5) and those with low/none (0-1), with adjustment for age and sex, and then cardiovascular disease (CVD) risk factors.

Results: 42,079, participants without prior atherosclerotic CVD had images for ML-AAC24 assessment including 22,258 (52.9%) women, 15,752 (37.4%) reported ever smoking and a mean age of 63.9±7.7 years. There were 1032 AF/F events reported (rate=2.41%). After adjustment for traditional risk factors AAC, was associated with an approximate 29% increased risk of AF/F (95%CI=1.03 to 1.62) only in those with severe AAC. However, after inclusion of lipids in the final model, this lost significance with only a marginal change in the effect estimate suggestive of a clinically important effect being observed.

Interpretation: Only the presence of severe AAC was associated with a clinically important increased risk of AF/F. Therefore, AAC captured at the time of bone density testing may have application in identifying individuals at risk of AF/F and possibly explains the excess cardiovascular burden in patients with osteoporosis given the robust inverse relationship between bone density and AAC.

669b008bc0551-Screen+Shot+2024-07-20+at+10.14.54+am.png

  1. 1. Sharif N, Gilani SZ, Suter D, Reid S, Szulc P, Kimelman D, Monchka BA, Jozani MJ, Hodgson JM, Sim M, Zhu K, Harvey NC, Kiel DP, Prince RL, Schousboe JT, Leslie WD, Lewis JR. Machine learning for abdominal aortic calcification assessment from bone density machine-derived lateral spine images. EBioMedicine. 2023 Aug;94:104676. doi: 10.1016/j.ebiom.2023.104676. Epub 2023 Jul 11. PMID: 37442671; PMCID: PMC10435763.