Reference values for skeletal muscle mass and fat mass measured by bioelectrical impedance in 390 565 UK adults

BACKGROUND: Loss of skeletal muscle mass (SMM) increases the risk of frailty and, together with excess fat mass (FM), is a risk factor for cardio‐metabolic disease. However, use of body composition measurements in nutritional surveillance and routine clinical practice is limited by the lack of reference data. Our aim was to produce age‐specific and sex‐specific reference values for SMM and FM in the White ethnic adult population in the UK. Secondary objectives were to examine the tracking over time using a subsample of the population with repeated measures of body composition and to assess the validity of these reference values in different ethnic subgroups. METHODS: We used data from segmental bioelectrical impedance analysis (BIA) in 390 565 participants, aged 40–69 years, in the UK Biobank, and data from dual‐energy X‐ray absorptiometry from n = 905 participants to validate the BIA measurements. SMM was calculated as the sum of the predicted muscle mass from the limbs. The LMS method was used to produce percentile curves for the SMM index (SMMI = SMM/height(2)) and the FM index (FMI = FM/height(2)). We investigated the validity of the White ethnic reference values by plotting z‐scores (99.7% confidence interval) from Black and Asian groups to check if their confidence interval included zero. Longitudinal trajectories were predicted based on the baseline z‐scores and the correlation between repeated measurements at follow‐up. RESULTS: The percentile curves show that SMMI declines in men from the age of 40, whereas in women, SMMI is more stable and decreases only slightly among women in the higher percentiles. FMI increases with age in both men and women. Women have higher FMI and lower SMMI than men in all age groups. The validity of the White‐based reference values for non‐White ethnic groups is poor. Longitudinal trajectories in body composition in the subsample of participants with a follow‐up assessment show regression towards the mean in both men and women, with some evidence of declining SMMI only among men. The predicted 90% limits for the expected 5 year trajectories of SMMI and FMI can be used to identify people with unusual trajectories and in clinical practice to identify and track individuals at risk of excessive loss of SMM. CONCLUSIONS: These body composition reference values developed from BIA in a middle/older‐aged healthy White ethnic population in the UK could be used as a simple assessment tool for nutritional surveillance and to identify individuals with low SMMI or high FMI who may be at increased risk of disease and/or frailty.