Machine learning derived abdominal aortic calcification is associated with physical frailty in community-dwelling adults: the UK Biobank Imaging Study.
Authors
Affiliations (17)
Affiliations (17)
- Nutrition & Health Innovation Research Institute, School of Medical and Health Sciences, Edith Cowan University, Perth, WA, Australia.
- Nutrition & Health Innovation Research Institute, School of Medical and Health Sciences, Edith Cowan University, Perth, WA, Australia. [email protected].
- School of Human Sciences, The University of Western Australia, Perth, WA, Australia. [email protected].
- Menzies Institute for Medical Research, University of Tasmania, Hobart, Australia.
- Applied Health Research Unit, Nuffield Department of Population Health, University of Oxford, Oxford, UK.
- Centre for Artificial Intelligence & Machine Learning, School of Science, Edith Cowan University, Perth, WA, Australia.
- Department of Computer Science and Software Engineering, The University of Western Australia, Perth, WA, Australia.
- Mater Research Institute, The University of Queensland, Translational Research Institute, Woolloongabba, QLD, Australia.
- Medical School, The University of Western Australia, Perth, WA, Australia.
- McMaster Institute for Research On Aging, McMaster University, Hamilton, ON, Canada.
- Department of Health Research Methods, Evidence, and Impact, McMaster University, Hamilton, ON, Canada.
- Labarge Centre for Mobility in Aging, McMaster University, Hamilton, ON, Canada.
- Department of Internal Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Canada.
- Park Nicollet Clinic and HealthPartners Institute, HealthPartners, Minneapolis, USA.
- Division of Health Policy and Management, University of Minnesota, Minneapolis, USA.
- MRC Life Course Epidemiology Centre, University of Southampton, Southampton, UK.
- NIHR Southampton Biomedical Research Centre, University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, UK.
Abstract
Clinical cardiovascular disease (CVD) is often present in frail individuals. However, it remains unclear whether subclinical CVD, e.g., abdominal aortic calcification (AAC), is associated with frailty. This study investigated the cross-sectional relationship between AAC scored using a validated machine learning model (ML-AAC24) and physical frailty. 49,081 participants from the UK Biobank Imaging Study without atherosclerotic CVD (ASCVD) diagnosis were included. ML-AAC24 extent was categorised as low, moderate and high, based on established severity categories. Physical frailty was based on a modified Fried's frailty phenotype comprising weak hand grip strength, slow walking speed, weight loss, exhaustion, and physical inactivity. Individuals with three or more deficits were considered frail, while one or two deficits was considered pre-frail. Multivariable-adjusted multinominal logistic regression models were used to test the associations between ML-AAC24 extent and frailty status. One in five individuals had moderate or high ML-AAC24. Compared to individuals with low ML-AAC24, those with moderate and high ML-AAC24 had greater odds of being pre-frail (ORs 1.06 95%CI 1.00-1.12 moderate; 1.14 95%CI 1.04-1.26 high) or frail (ORs 1.27 95%CI 1.12-1.44 moderate; 1.58 95%CI 1.31-1.91 high), adjusted for multiple covariates. When stratified by sex, similar results for frailty were recorded. In a population, those with moderate and high ML-AAC24 were more likely to present as pre-frail and frail. AAC identified from lateral spine images obtained during routine bone density testing, could serve as a useful marker for the early detection of frailty, highlighting the importance of multimodality care.