Associations of plasma biomarkers of Alzheimer's disease pathology with modifiable risk factors and cognitive and functional outcomes: evidence from cross-sectional prediction and latent path analyses

Biological ageing clocks integrate multiple biomarkers—such as epigenetic and proteomic measures—to capture biologically meaningful variation in ageing that goes beyond chronological age and helps predict the risk of age‐related diseases and mortality [, , ]. The search for biomarkers remains a key route in the field of GeroScience, aiming to prevent or delay the onset of diseases. However, current biological ageing clocks have centred on blood biomarkers, which rely on high‐throughput assays from blood samples. An individual's anthropometry can also resemble the appearance of ageing and may track the pace of ageing over time. Visual perceptions of someone who ‘looks younger or older than their chronological age’ may be associated with biological ageing. Recently, anthropometric‐based ageing clocks, using waist circumference and waist‐to‐hip ratio, have shown associations with body composition and mortality [, ]. But these initial studies are limited to a few anthropometric measures, which may lack comprehensive and unique anthropometric characteristics as they age. A digital 3D body scanner can precisely assess multidimensional anthropometric measures simultaneously and provide valid information on body volumes and body composition [, , ]. In this study, we aimed to use machine learning to identify age‐related features from whole‐body multidimensional anthropometric measures via a 3D body scanner. Further, we aimed to examine whether accelerated anthropometric ageing, indicated by a greater gap between anthropometric‐predicted age and the chronological age, would be associated with longitudinal changes in ageing phenotypes and function across the adult lifespan beyond the chronological age.