Abstract
Skeletal muscle echo intensity (EI) is affected by ageing and physical activity; however, the effects of nutrition are less understood. The aim of this study was to explore whether habitual nutrient intake may be associated with ultrasound-derived EI. Partial least squares regression (PLSR) models were trained on an initial sample (n=100, M=45; F=55; 38±15 years) to predict EI of two quadriceps muscles from 19 variables, using the 'jack-knife' function within the 'pls' package (RStudio), which was then tested in an additional dataset (n= 30, M=13; F=17; 38±16 years). EI was determined using B-mode ultrasonography of the rectus femoris (RF) and vastus lateralis (VL) and nutritional intake determined via three-day weighed food diaries. Mean daily intake of specific nutrients were included as predictor variables with age, sex and self-reported physical activity. PLSR training model 1 explained ~52% and model 2 ~46% of the variance in RF and VL EI, respectively. Model 1 also explained ~35% and model 2 ~30% of the variance in RF and VL EI in the additional testing dataset. Age and biological sex were associated with EI in both models (P<0.025). Dietary protein (RF: β=-7.617,VL: β=-7.480), and selenium (RF: β=-7.144,VL: β=-4.775) were associated with EI in both muscles (P<0.05), whereas fibre intake (RF: β=-5.215) was associated with RF EI only and omega-3 fatty acids (n-3/ω-3 FAs, RF: β=3.145) with VL EI only (P<0.05). Therefore, absolute protein, selenium, fibre and n-3 FAs may be associated with skeletal muscle EI, although further mechanistic work is required before claiming causal inference.Skeletal muscle echo intensity (EI) is affected by ageing and physical activity; however, the effects of nutrition are less understood. The aim of this study was to explore whether habitual nutrient intake may be associated with ultrasound-derived EI. Partial least squares regression (PLSR) models were trained on an initial sample (n=100, M=45; F=55; 38±15 years) to predict EI of two quadriceps muscles from 19 variables, using the 'jack-knife' function within the 'pls' package (RStudio), which was then tested in an additional dataset (n= 30, M=13; F=17; 38±16 years). EI was determined using B-mode ultrasonography of the rectus femoris (RF) and vastus lateralis (VL) and nutritional intake determined via three-day weighed food diaries. Mean daily intake of specific nutrients were included as predictor variables with age, sex and self-reported physical activity. PLSR training model 1 explained ~52% and model 2 ~46% of the variance in RF and VL EI, respectively. Model 1 also explained ~35% and model 2 ~30% of the variance in RF and VL EI in the additional testing dataset. Age and biological sex were associated with EI in both models (P<0.025). Dietary protein (RF: β=-7.617,VL: β=-7.480), and selenium (RF: β=-7.144,VL: β=-4.775) were associated with EI in both muscles (P<0.05), whereas fibre intake (RF: β=-5.215) was associated with RF EI only and omega-3 fatty acids (n-3/ω-3 FAs, RF: β=3.145) with VL EI only (P<0.05). Therefore, absolute protein, selenium, fibre and n-3 FAs may be associated with skeletal muscle EI, although further mechanistic work is required before claiming causal inference.