Muscle mass index and animal source of dietary protein are positively associated with insulin resistance in participants of the NuAge study

Objectives Contribute evidence towards the complex interrelationships of body composition, insulin sensitivity and protein intake independently from adiposity in an older population. Design This is a cross-sectional analysis of an existing dataset in which a literature-supported model linking together the variables of interest is tested using path analysis. Setting The loss of muscle mass has been implicated in the development of insulin resistance. We propose to test associations of muscle mass with insulin sensitivity and their respective associations with animal and vegetable sources of protein intake, independently from adiposity. Participants Non-diabetic participants aged 68–82 years from the NuAge study with all available measures (n=441) were included. Measurements A model considering age, sex, chronic diseases, physical activity; smoking and sources of protein intake influencing body composition components and insulin sensitivity was created and tested with Path Analysis for their independent associations. Muscle mass index (MMI; kg/height in m 2 ) and % body fat were derived from DXA and BIA. Insulin resistance was estimated by the Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) score and physical activity by the Physical Activity Scale for the Elderly (PASE) questionnaire. Protein intakes were obtained from three non-consecutive 24h-diet recalls. Results In the final model, direct positive associations were observed between HOMA-IR score and MMI (β=0.42; 95%CI: 0.24; 0.6) and % body fat (β=0.094; 95%CI: 0.07; 0.11). There were no direct associations between animal protein intake and MMI or with HOMA-IR. There was a significant direct negative association between plant protein intake and MMI (β= -0.068; 95%CI: -0.13; -0.003) and significant indirect associations mediated through MMI and % body fat between HOMA-IR and animal protein intake (β=0.0321; 95%CI: 0.01; 0.05), as well as plant protein intake (β= -0.07; 95%CI: -0.1; 0.0). Conclusions Our final model indicated that MMI and HOMA score were significantly positively associated. Protein intake sources were related to HOMA-IR score differently through MMI and % body fat, respectively.